Governing with Artificial Intelligence

AI in justice administration and access to justice

Digital technologies and data have significant potential to support access to justice, as well as to support its resilience, efficiency, effectiveness and fairness, in line with the OECD Recommendation on Access to Justice and People-Centred Justice Systems (2023[289]). Of all the technologies driving this transformation, AI stands out as both remarkable and novel.

So far, AI is unevenly impacting justice systems. While certain countries have embraced AI for practical applications within their justice sectors, others are still in the early stages of integrating more basic technological advancements, let alone AI. AI applications in this area range from automating routine administrative tasks related to case management to offering predictive analytics and supporting legal research. Often challenged by backlogs, lengthy procedural requirements and limited accessibility, the justice system is set to gain immensely from AI integration. AI has great potential to improve efficiency and responsiveness, enhance the delivery of services tailored to people’s needs, support routine tasks and expand capabilities across a variety of justice domains (Redden and O’Donovan Dix, 2020[290]; CEPEJ, 2018[291]).

Improper development or use of AI in the justice sector can result in disparate treatment between individuals or groups. Governance frameworks and guardrails to assess risks and incidents, and remedies for those affected by AI systems can help ensure the trustworthy use of AI in justice.

Current state of play

Making internal operations more efficient

AI's impact on productivity and responsiveness in the justice system can be transformative. Advanced AI algorithms can analyse incoming cases, categorise them based on complexity or urgency, and assign them to appropriate departments or judges (Reiling, 2020[292]). This can significantly reduce administrative bottlenecks, help ensure more efficient use of resources and enhance effectiveness in the delivery of justice. For instance:

• The French Court of Cassation (2024[293]) developed an AI system that categorises and routes incoming appeals and petitions to the appropriate chambers based on training from past cases, making case management more efficient and ensuring cases the appropriate level of scrutiny.

• In Brazil, the VICTOR AI system automates the examination of appeals to the Supreme Court by identifying cases with “general repercussions”, a requirement for processing appeals (Supreme Court of Brazil, 2024[294]). While a court clerk takes 44 minutes to evaluate whether an appeal meets conditions to move forward, VICTOR AI spends a couple of seconds. In addition, Brazil's Superior Council of Labour Justice launched Chat-JT, a generative AI tool that assists judges, court staff and interns by automating tasks (e.g. legal research, document analysis, the drafting of standardised summaries). Launched in February 2025, the platform also allows users to create personalised assistants tailored to specific needs within the judicial workflow. By integrating with other internal systems, Chat-JT provides advanced search capabilities across internal databases, aiming to streamline daily operations and improve decision-making processes within the Labour Justice (CNJ, 2025[295]).

• Automated transcription can help improve access to court proceedings. In Slovenia, for instance, automated transcription is improving access to court proceedings (Slovenia Ministry of Justice, 2024[296]). This technology rapidly converts spoken words into written text, ensuring greater transparency and allowing more people to review and understand legal proceedings (CMS, 2023[297]). In Spain, transcription services, referred to as “textualisation”, have been operating since 2019, integrated into the HORUS judicial electronic file viewer. This combination of systems not only provides transcripts but also enables rapid navigation of hearing recordings. It does so by 1) associating each segment of the textualisation to the corresponding time frame of the recording, and 2) identifying the different voices featured in the recording, allowing users to track which parties intervene at any given moment. This is particularly important in Spain, where the recording itself serves as the official record of hearings.

• Colombia’s PretorIA system assists the Constitutional Court in managing the high volume of tutela cases, or legal actions filed to protect human rights. With over 600 000 custodianships (“tutelas”) received annually, PretorIA helps pre-select cases for review by detecting predefined legal criteria. Ensuring human oversight and explainability has been a key priority, and PretorIA was redesigned in 2020 to use interpretable topic modelling rather than opaque neural networks. Judges remain the ultimate decision-makers, maintaining human authority over case selection (OECD/CAF, 2022[298]).

• In Greece, AI implementation facilitates judicial processes via automated document processing aligned with national legal frameworks and the European AI strategy. NLP technologies analyse legal documents, enabling efficient case law searches and court record management.1 Other current initiatives integrate AI for speech-to-text conversion, document translation, and witness statement interpretation. Additional applications include forensic services support and judicial administration optimization, all maintained under human oversight to ensure human rights protection while improving efficiency.2

Generative AI systems are also helping to enhance efficiency in justice systems. Automated transcription – one of the more basic forms of generative AI – can improve access to court proceedings, as is being done in Slovenia and Spain. This technology rapidly converts spoken words into written text, ensuring greater transparency and allowing more people to review and understand legal proceedings.

NLP and generative AI have also supported improving case management and processing of judicial documents. In the case of Spain, by automating repetitive tasks and improving metadata extraction, AI-enabled services have streamlined daily workflows for justice system personnel, enabling more effective allocation of human resources and allowing staff to focus on higher-value activities such as complex legal analysis (Box 5.61).3

Box 5.61. Enhancing document management in Spain

Spain’s Ministry of the Presidency, Justice and Parliamentary Relations has developed an in-house suite of NLP and generative AI tools to support efficient management and processing of judicial documents. These tools assist legal professionals in classifying, analysing, summarising and anonymising court-related texts, while ensuring compliance with data protection standards.

The classification functionality allows documents to be sorted by jurisdiction, registry type and document category (e.g. legal submissions, notifications), thereby improving information retrieval and systematisation across various branches of the justice system. The suite also includes advanced analytical features that extract and visualise named entities, key dates, monetary values and relationships between entities. In addition, the system performs semantic similarity detection, helping to harmonise inconsistent name variants across documents.

Summarisation capabilities are tailored to various legal document types — such as rulings and orders — and offer four distinct modes, ranging from extractive to generative, including a “legal to plain language” option designed to enhance accessibility.

The suite is available through three channels:

1. a secure online portal for use by Justice personnel

2. the Justice Folder, Spain’s one-stop-shop for citizens where the anonymisation and summarisation tools are also accessible

3. integration with other digital justice products.

A notable example of this integration is LexNET, Spain’s electronic information exchange platform connecting judicial bodies with legal professionals, law enforcement agencies, hospitals and other stakeholders. Within LexNET, NLP scans documents to identify and extract named entities, which are then used to auto-populate required form fields that previously had to be filled in manually. This automation significantly reduces processing time from several minutes to a matter of seconds.

Future enhancements to the NLP and generative AI suite may include translation capabilities to ensure bilingual access across Spain’s official languages (Spanish, Catalan, Galician, Valencian and Basque).

Source: Information provided by officials of the Ministry of Presidency, Justice and Parliamentary Relations to the OECD.

More sophisticated generative AI systems are helping to create content and perform information analyses. MARIA (acronym for “AI Writing Support Module”) is an LLM-driven tool that supports the Brazilian Supreme Court in summarising documents, producing first drafts of reports and conducting preliminary assessments to identify incoming cases with judicial precedents (STF, 2024[299]). MARIA’s implementation brings important efficiency gains; the automation of repetitive tasks frees up court clerks’ time for more complex activities. Similarly, Argentina is introducing LLM to complement its current stack of AI tools to support the judicial sector (Box 5.62).

Box 5.62. Prometea and ChatGPT to make Argentina’s judicial sector more efficient

Prometea is an AI-powered system developed in 2017 by the Public Prosecutor’s Office of the Autonomous City of Buenos Aires and the Innovation and Artificial Intelligence Laboratory (IALAB) of the University of Buenos Aires. Designed to automate repetitive judicial tasks and expedite case proceedings, Prometea acted as a virtual assistant, predicting case solutions based on previous rulings and helping assemble case files. By significantly reducing processing times, it allowed judicial officials to focus on more complex cases requiring human judgment. Between 2017 and 2020, Prometea was instrumental in resolving 658 cases related to housing rights, labour disputes and disability rights. The system achieved a 90% accuracy rate in its predictions, with prosecutors’ decisions aligning with the AI’s recommendations in most cases. By enhancing workflow efficiency, Prometea increased productivity by nearly 300%, enabling legal professionals to process around 490 cases per month compared to 130 before its implementation.

Over time, the Inter-American Court of Human Rights and other legal bodies adopted Prometea. However, its expansion to new case categories required extensive retraining and algorithmic adjustments, leading to concerns about its scalability. In 2024, developers began phasing out Prometea in favour of generative AI systems, including ChatGPT, for certain legal tasks. Since May 2024, the Public Prosecution Service of the City of Buenos Aires has used ChatGPT to predict rulings for some public employment cases related to salary demands. Justice employees have uploaded case documents into ChatGPT, which analyses patterns, offers a preliminary classification from a catalogue of templates, and drafts decisions. So far, ChatGPT has been used for 20 legal rulings, reducing the time needed to draft a ruling from an hour to about 10 minutes.

Source: (OECD/CAF, 2022[298]), https://restofworld.org/2024/buenos-aires-courts-adopt-chatgpt-draft-rulings.

The time courts take to provide or decline protection measures for people in vulnerable situations, such as domestic violence, can have severe consequences for victims. Swift judicial responses are essential to ensure the safety and well-being of individuals facing immediate threats. Prolonged waiting periods for protective orders can leave victims vulnerable to further abuse. Some countries have implemented solutions using a mix of methods, including generative AI, to help provide timely judicial responses and, when applicable, issue protective measures4 to uphold victims’ rights (Box 5.63).

Box 5.63. Redesigning justice services with AI for effective protection measures in Peru

Peru's Amauta Pro system, developed in-house by the Superior Court of Justice of Lima Norte, has been piloted with domestic violence cases. It uses an AI system and standardised templates of judicial decisions to perform various tasks essential for the determination of the resolution to provide or decline protection measures. The automated tasks are:

• Conversion and categorisation: The system converts machine-readable files to text, categorises each page based on content and verifies the validity of file codes. This categorisation helps in organising and managing case files more efficiently.

• Case matching: Amauta Pro searches for open proceedings involving the same parties or related facts. If matches are found, the system prepares draft resolutions, streamlining the process of case consolidation.

• Data extraction and cleaning: For cases without matches, the AI system extracts data from police reports and other relevant documents, cleans the text and identifies key information.

• Drafting resolutions: The system drafts resolutions based on extracted and analysed data, which judges can review and finalise. This reduces the time required to draft resolutions.

The project aims to improve access to justice for victims of violence by providing timely and effective protection measures. The system reduces the time taken to draft a resolution from 3 hours to 40 seconds. The project has garnered support from the highest levels of the Peruvian judiciary, including the President of the Judiciary, underscoring the project's importance and the confidence placed in its ability to transform the delivery of justice services in Peru. With over 800 court ruling proposals generated during its pilot phase, Amauta Pro enables judges to focus more of their time and resources on the critical aspects of decision-making rather than on time-consuming administrative tasks.

Source: (Government of Peru, 2023[300]).

AI algorithms can also be trained to identify and categorise data, including text containing personal information (e.g. name, address) from court documents and transcripts. This can be used to protect sensitive information and maintain privacy, critical aspects of handling justice cases. A number of country practices demonstrate the application of AI in this area (Box 5.64).

Box 5.64. AI-powered anonymisation of court decisions

Spain: Automated anonymisation tool for legal documents

Spain’s Ministry of the Presidency, Justice and Parliamentary Relations offers a robust anonymisation tool as part of its NLP and generative AI suite. The tool removes personally identifiable information from documents through three modes — automatic, initials-based and fully customisable — enabling the creation of shareable, GDPR-compliant documents and datasets. In addition to choosing the mode of anonymisation, users can specify the types of information they wish to anonymise. This includes data related to individuals (e.g. names), legal entities, addresses, identity documents (e.g. ID or passport numbers), as well as email addresses and social media profiles.

Croatia: Automated anonymisation of judicial decisions

As of January 2025, Croatia has operationalised ANON, an AI-powered tool designed to automate the anonymisation and publication of court decisions. Integrated with the national eSpis case management system, ANON automatically anonymises first-instance judgments and proposes indexing for publication on the publicly accessible portal, Search Engine for Decisions of the Courts of the Republic of Croatia (Tražilica odluka sudova Republike Hrvatske). For higher-instance courts, the system facilitates anonymisation, with final verification conducted by authorised court personnel. This initiative, supported by the National Recovery and Resilience Plan 2021–2026, aims to enhance judicial transparency, improve access to legal information and ensure compliance with personal data protection standards.

Austria: A combination of approaches to court decision anonymisation

Austria's Federal Ministry of Justice (BMJ) is implementing a comprehensive project to automate the anonymisation of court decisions. The project's primary goals are to improve efficiency, reduce manual effort in the anonymisation process, and enable wider publication and access to anonymised court decisions. This initiative uses a combination of AI approaches, including search-based methods using register data and regular expressions, NLP for named entity recognition, and custom ML algorithms to continuously refine the quality of results based on feedback.

France: AI-driven anonymisation for open data release

The French Court of Cassation’s AI-powered anonymisation engine is a key component in its efforts to increase transparency and accessibility of legal information. This system is designed to automatically identify and protect personal data within court decisions, preparing them for public release as part of an open data initiative.

Greece: AI-enabled summarisation and anonymisation

The Ministry of Justice in Greece is using AI to translate summarize and anonymize documents. Furthermore, the Ministry is introducing an AI assistant for providing instructions to citizens and lawyers regarding electronic and non-electronic services of the Greek Justice System.

Note: Register data are structured databases or official registries that contain names, case numbers or other identifying information. Regular expressions are pattern-matching rules used in text processing. They help detect and redact specific patterns, such as names, dates or addresses, based on predefined formats.

Source: (Spain Ministry of Justice, 2022[301]; Austria Federal Ministry of Justice, 2021[302]; Sommer, 2021[303]), Croatian Ministry of Justice officials, https://ministryofjustice.gr/?p=13929, https://ministryofjustice.gr/?p=15317.

Improving delivery of justice services

In OECD countries, AI is being deployed to support the delivery of and access to legal and judicial services by facilitating initial consultations through chatbots and virtual assistants. These AI systems can provide preliminary legal advice in accessible language, triage cases based on urgency and relevance, and guide users through court proceedings, as done in Portugal (Box 5.65).

Box 5.65. Portugal’s AI-powered chatbot to provide practical information

Portugal’s Ministry of Justice has implemented a Practical Guide to Justice (GPJ), an AI-powered chatbot that provides the public with easy access to justice-related information. The GPJ is trained to respond to various formulations of questions in an accessible manner. It provides answers based on information from the Digital Justice Platform, offering people practical guidance on justice-related matters. The chatbot is designed to speed up interactions between people and the justice system, helping them find the information they need more efficiently. The GPJ provides user-friendly, plain language responses to people's inquiries. It helps bridge the information gap between the complex legal system and the general public, making justice-related information more accessible to all.

Source: (Government of Portugal, 2024[304]).

Victim support services also stand to benefit from AI-driven innovations. These services, which often involve providing timely information, legal support and resources to victims of crime, can be augmented through the integration of AI technologies. AI-powered chatbots, for instance, can offer personalised support, providing victims with immediate access to information on their rights, available resources and steps to take following a crime (see example in Box 5.66).

Box 5.66. Chatbot supports child victims with tailored legal information

The i-ACCESS My Rights project implemented in Greece, Bulgaria and Romania provides age-appropriate and child-friendly legal information to children aged 13-18, particularly those who have been victims of abuse or crime. Offering general legal information and guidance on available support services, the i-ACCESS chatbot is designed to empower young victims by informing them of their rights within legal proceedings and directing them to relevant support networks and authorities. What sets this project apart is its strong emphasis on child participation in the development process through co-design sessions and Child Advisory Boards in each country. To ensure that the AI tool is truly responsive to the needs and preferences of its young users, children with lived experiences in the justice system actively contributed to the chatbot's design and functionality.

Source: https://justicewithchildren.org/en/i-access-my-rights.

Enabling predictions for a more proactive justice system

AI can enhance the justice system’s proactivity by enabling predictive analytics. This predictive approach allows justice systems to shift from a reactive to a proactive approach, focusing on prevention and informed decision-making. Two usual applications of AI in justice administration and access to justice are predictive policing – where AI analyses historical data to identify patterns and forecast potential events (see section on “AI in law enforcement and disaster risk management”) – and risk assessment tools – where AI is used to evaluate the likelihood of individuals becoming victims or re-engaging in similar criminal conduct in the future, supporting decisions on bail, sentencing and parole.

Some countries have been advancing in the implementation predictive policing and risk assessment tools, especially in the context of domestic violence and community violence prevention, with some promising implementations. Spain and the United States operate an extensive victim risk assessment system in the area of domestic violence (Box 5.67). Some other relevant examples on predictive policing are provided in this chapter’s section on “AI in law enforcement and disaster risk management”.

Box 5.67. Risk assessment system for domestic violence in Spain

Spain’s Comprehensive Monitoring System for Gender-Based Violence Cases (VioGén) is a risk assessment tool designed to enhance victim protection and prevent domestic violence escalations. It aims to strengthen the country’s capacity to respond to violence against women and girls using evidence-based assessments. Since its introduction in 2007, VioGén has processed over 670 000 cases, facilitating a coordinated response across the judiciary, national and local police forces. The system operates through a structured risk assessment process, starting when a victim reports abuse or when police intervene in a domestic incident. Officers complete a detailed questionnaire, assessing factors such as the aggressor’s history, the severity of prior violence and victim vulnerabilities. The algorithm then assigns a risk level — low, medium, high or very high — determining the necessary protective measures, such as police check-ins or emergency alarms. Cases deemed “of special relevance,” based on risk levels, can be flagged for prosecutors and judges. Regular follow-up assessments adjust risk classifications, ensuring responses remain appropriate over time.

Source: (Government of Spain, 2007[305]; Algorithm Watch, 2020[306]).

These tools help identify patterns in victimisation and anticipate and address potential issues before they escalate. Yet they also require careful implementation to mitigate biases and uphold ethical standards within the justice system. In Spain, while evidence suggests that the VioGén risk assessment tool (Box 5.67) strengthens victim protection, concerns — regarding algorithmic “black box”, risks of potential misclassification and limited literacy on the use of digital tools and approaches sensitive to women and girls — underscore the importance of appropriate training for users, and of transparency and oversight mechanisms. VioGén’s algorithm seems to remain largely opaque; its exact formula is undisclosed, raising questions about accountability and fairness. Reports also indicate that in 95% of cases, officers adhere to VioGén’s risk classification without modification, despite having the discretion to override it (Algorithm Watch, 2020[306]).

Another example, in the United States, is the Correctional Offender Management Profiling for Alternative Sanctions (COMPAS), a proprietary risk assessment tool that evaluates the likelihood of a defendant reoffending. Used in several U.S. jurisdictions, COMPAS was designed to support courts in making informed decisions regarding bail, sentencing and parole by predicting general and violent recidivism risks (Equivant, 2025[307]). The system analyses data from a comprehensive questionnaire that covers factors such as criminal history, substance abuse, social connections and personal demographics. This information is processed through a proprietary algorithm to classify individuals as having a low, medium or high risk of reoffending. These risk scores assist legal professionals tailor interventions and supervision levels to each offender's assessed risk.

While COMPAS has brought efficient gains — in processing time and standardisation across jurisdictions — the system was thrust into the spotlight due to its 68% accuracy level (De Miguel Beriain, 2018[308]; Grgić-Hlača et al., 2018[309]). As in the case of VioGén, the limited algorithmic transparency of COMPAS also contributed to ethical questions over whether it was a trustworthy system to take fundamental decisions such as defendants’ freedom. Some jurisdictions have since sought more transparent alternatives — for example, the Public Safety Assessment (PSA), a non-proprietary statistical risk assessment method (APPR, 2025[310]) — while others continue using tools like COMPAS under closer scrutiny.

AI has enormous potential to improve proactiveness in justice systems (see “Untapped potential and way forward” below). Yet, there is still a long way before reaping the benefits of AI predictive systems. This includes expanding training to justice professionals on the use of AI tools, greater algorithmic transparency, and human oversight (see “Managing risks and challenges” below).

Monitoring judicial performance and enhancing accountability

AI has proven effective in various scenarios to enhance the transparency and accountability of government activities (OECD, 2022[152]). AI applications for external and internal oversight in the justice sector focus on performance monitoring of judicial systems and enhancing accountability. AI systems can provide insights into court performance, identify potential inconsistencies in decision-making and support policy evaluation. These systems should be accompanied by pre-existing governance mechanisms on checks and balances (e.g. administrative independence and autonomous judicial review) to ensure the judiciary’s autonomy. AI-driven court performance monitoring could risk undue interference if there are no protections in place. Checks and balances (such as independent oversight bodies, administrative independence and autonomous judicial review) help prevent AI tools from being used to influence case outcomes or unfairly evaluate judges. AI can highlight patterns in judicial decisions, but without independent governance mechanisms, these insights could be potentially misused (e.g. penalising judges for unpopular rulings instead of focusing on systemic improvements).

While less common than other applications, these oversight activities represent an important area with potential for AI development in justice systems. The RAFA 2030 project implemented by the Brazilian Supreme Court (STF) stands out as an innovative example of AI use for oversight (Box 5.68).

Box 5.68. Harnessing AI oversight in Brazil’s Supreme Court

The RAFA 2030 project implemented by the Brazilian Supreme Court (STF) uses ML and deep learning (DL) to classify data by category, by processes that are on the STF’s plenary agenda, by origin of the process, by class, by completed or ongoing processes, and by Sustainable Development Goals (SDGs) 2030. RAFA 2030 not only supports internal performance evaluation but also facilitates external oversight by providing a standardised framework for assessing the implementation of the SDGs agenda in the context of Supreme Court decisions. Automated classification of cases with SDGs provides a clear overview of how the court's decision aligns with the 2030 Agenda.

In addition, by aligning court cases with global sustainability objectives, RAFA 2030 enables the comparison of the STF's performance with other international higher courts. The system has already supported the classification of 3 315 court cases, demonstrating its potential to enhance transparency and align judicial activities with broader societal goals.

Source: https://portal.stf.jus.br/hotsites/agenda-2030.

In Israel, the Israeli Ministry of Justice has implemented an AI-driven initiative within the Public Defender’s Office to support oversight in Office, to enhance the quality of legal representation and strengthen supervisory mechanisms. The project integrates ML, NLP and optical character recognition (OCR) technologies to streamline the review and management of criminal case files. This initiative supports the oversight of defence quality by enabling automated analysis of legal documents, improving consistency checks and identifying potential issues in case handling.5

Evidence of impact

Pioneering initiatives like those discussed above have made legal information more accessible, breaking barriers that have long kept many people from fully engaging with the justice system. These AI-driven tools help address the needs of many people — such as identifying the most suitable justice services for their individual situation — promote a more informed citizenry and pave the way for more effective and efficient justice systems.

One of AI’s most significant impacts has been its ability to accelerate various legal processes. By reducing the time needed for document review and categorisation, AI allows more cases to be handled in less time. The French Court of Cassation’s aforementioned AI solution for pre-directing appeals has, for example, obtained results on the supplementary briefs from 2020, showing that 87% are of good pre-direction (Court of Cassation of France, 2024[293]). As a case improving judicial responsiveness, Peru's Amauta Pro AI system (Box 5.63) has transformed the speed at which courts can respond to victims of violence, reducing the time needed to draft resolutions for protection measures from three hours to 40 seconds. It is also fundamentally enhancing the judicial system’s overall responsiveness in addressing cases of violence against women and family members, contributing to effective access to justice.

While results and impact data exist for some specific use cases, as discussed in the examples above, governments have not yet extensively collected or documented results and impacts on many others. Despite growing interest in AI applications across justice systems, countries have not established systematic mechanisms to collect and analyse data on their impact. Existing evidence remains fragmented, often limited to pilot studies or sector-specific evaluations, with few comprehensive assessments at either the organisational or national level (Queen Mary University of London; Centre for European Policy Studies, 2021[311]). This evidence gap at organisational and national levels means reformers are not learning from experience. Potentially effective AI innovations might be overlooked or, conversely, unproven tools scaled up inappropriately.

While some jurisdictions report efficiency gains from AI-assisted case management or automated legal research, there is little empirical data on whether these tools improve the delivery of people-centred justice or enhance decision-making accuracy. Without structured monitoring, governments risk deploying AI solutions without clear evidence of their benefits or unintended consequences, limiting their ability to make informed policy decisions. The lack of evidence of impact reflects a broader systemic issue: justice systems have often encountered challenges in collecting and effectively using data (World Justice Project, 2021[312]) Historically, justice systems have relied heavily on paper-based processes (Byrom, Piccinin-Barbieri and Wells, 2024[313]) (see “Managing risks and challenges” below).

Managing risks and challenges

Associated risks

• Inadequate or skewed data in AI systems

• Lack of transparency and explainability

• Inaccuracy, falsehoods, and unreliability

• Over-reliance on AI and resistance to government use of AI

If AI systems rely upon inadequate or skewed data, it could lead to inaccurate or adverse outcomes for some individuals or groups. With regard to justice administration and access to justice, this could result in, for instance, inaccurate risk assessments (e.g. for likelihood of recidivism) that could result in lead to adverse impacts for some individuals.

The lack of transparency of AI systems used for justice administration was identified as another concern in the review of use cases. Taking the examples mentioned in “Enabling predictions for a more proactive justice system”, while were designed to enhance efficiency and decision-making, AI systems such as VioGén and COMPAS often operate as black boxes, with their underlying methodologies, weighting of factors and potential biases shielded from public scrutiny. The proprietary nature of certain AI systems or lack of external audits prevents defendants from understanding or challenging the risk scores that influence decisions, raising concerns about due process and fairness in judicial decisions. While lack of algorithmic transparency is not an uncommon issue, it has a very sensitive dimension in the administration and access to justice as it can be detrimental to human rights and fundamental freedoms and reinforce systemic inequalities in societies. Lack of algorithmic transparency has important impacts on explainability of judicial decisions. When AI systems are not transparent, it becomes challenging to understand the rationale of decisions. This opacity can result in difficulties in challenging and ensuring fairness of decisions, leading to distrust on the use of the tools and ultimately on justice systems.

The reliance on AI, and potential concerns regarding over-reliance on AI, for accountability introduces new layers of complexity that require robust oversight mechanisms. Not all judges are convinced of using AI systems in the justice systems and have concerns on their applications. For example, 27% of respondents to a UNESCO survey among judicial operators expressed concerns about the AI chatbot’s output quality, highlighting issues such as inaccuracy, falsehoods, and unreliability. Governments can mitigate these risks by implementing strong policy and regulatory frameworks. Spain, for instance, established a framework for the responsible and ethical use of AI in justice, recognising AI’s potential to support, but not substitute jurisdictional decision-making (Box 5.69). Greece’s Permanent Scientific Committee of the Ministry of Justice, with the aim of examining the impacts of the introduction of AI in the judicial system, has drawn up guidelines that must be followed by the competent bodies of the Courts and Prosecutor's Offices.6

Box 5.69. Spain: National Policy on the Use of AI in the Administration of Justice

In June 2024, Spain’s Technological Committee for the Electronic Judicial Administration (CTEAJE) approved the National Policy on the Use of Artificial Intelligence in the Administration of Justice. This policy establishes a foundational framework for the responsible, lawful, and ethical use of AI within judicial settings. Grounded in the provisions of the Spain’s Justice Digital Efficiency Act (Royal Decree-Law 6/2023), which regulates the use of AI in the justice sector, the policy seeks to balance technological innovation with judicial integrity. It recognises AI’s potential to support, but not substitute, jurisdictional decision-making.

The policy applies to any AI system that handles judicial data, from the initial filing of a case to its final resolution. It draws a clear line between systems that could affect judicial independence and those intended purely for administrative purposes. The guidelines are binding for all personnel operating within the justice system, including external collaborators in both the public and private sectors who have access to judicial data. The policy was agreed upon by key institutional stakeholders of Spain’s Justice system: the General Council of the Judiciary (CGPJ), the Office of the Attorney General, all regional administrations with competence over the Justice Administration, and the Ministry of the Presidency, Justice and Parliamentary Relations.

In addition to these principles, the policy provides operational guidance, outlining acceptable uses of AI — such as internal document summarisation and automation of non-sensitive administrative tasks — while explicitly prohibiting others. Prohibited uses include automated decision-making without human oversight and the generation of binding legal content based on protected data. The policy also requires clear labelling of AI-generated outputs and affirms national sovereignty over the data and algorithms deployed.

Oversight responsibilities are clearly defined in the policy. When AI tools have an impact on judicial functions, the CGPJ is responsible for conducting algorithmic audits to safeguard judicial independence and ensure transparency. For tools used in non-judicial contexts, oversight is assigned to CTEAJE or the relevant administrative authority. Additional mechanisms include the maintenance of FAT (Fairness, Accuracy, and Transparency) registers for AI systems and sustained inter-institutional collaboration to promote ethical innovation throughout Spain’s judicial ecosystem.

Source: (Spain Ministry of Justice, 2024[314]); Information provided by officials of the Ministry of Presidency, Justice and Parliamentary Relations to the OECD.

In addition to algorithmic transparency, countries should consider implementing effective human oversight in the use of AI tools in their justice systems, including AI systems with clear decision-making processes allowing stakeholders to understand capabilities and limitations and interpret outcomes, conduct ongoing tracking of AI performance, and allow human intervention and control (OECD, 2024[91]). Ethical guidelines, such as the 2018 European Ethical Charter on AI in Judicial Systems, was a first step in this regard, outlining principles to ensure that AI development in judicial systems adheres to fundamental rights as conceptualised under EU law (CEPEJ, 2018[291]). Continuous monitoring and improvement should be built into an operational framework, allowing for regular audits of AI performance and outcomes. Additionally, operational plans should account for the evolving nature of law and help ensure that AI systems can be updated to reflect changes in legislation or precedent.

Implementation challenges

• Lack of high-quality data and the ability to share it

• Skills gaps

• Lack of actionable frameworks and guidance on AI usage

Historically, justice systems were predominantly paper based and, until recently, there was limited attempt to implement sound data governance in justice systems (Byrom, Piccinin-Barbieri and Wells, 2024[313]). This legacy is still very present today. For example, according to the OECD OURdata Index, countries perform relatively better in making statistics and geospatial data available, whereas weaker performance include data related to justice (OECD, 2023[315]). In addition, where data are collected, they are often held and stored in formats that are not easily accessible (e.g. in narrative format, PDF or paper files), or recorded inconsistently across agencies (Byrom, Piccinin-Barbieri and Wells, 2024[313]). The justice system deals with vast amounts of complex, nuanced data, including case law, statutes and individual case details. Ensuring this justice data is accurate, comprehensive and free from historical biases is paramount. This implies improving the quality and representativeness of data used to train AI systems, the robustness of systems housing this data, and the ability to handle sensitive information securely (Byrom, Piccinin-Barbieri and Wells, 2024[313]). Moreover, the technical infrastructure should be capable of processing this data efficiently while maintaining the highest standards of security and privacy.

Operational suitability focuses on integrating AI seamlessly into existing processes and workflows while enhancing digital competencies. This encompasses providing training to legal professionals, including judges, to effectively use AI tools and interpret AI-assisted insights, and adapting processes that need to be optimised to incorporate AI-driven solutions. Some countries have taking steps in this regard as AI tools are launched (CNJ, 2025[295]).

The use of AI raises questions of regulatory compliance, which is particularly important in the justice system due to the high stakes involved in judicial decisions and the fundamental or constitutional rights at play. The EU AI Act exemplifies the evolving regulatory landscape, classifying AI systems in law enforcement and administration of justice as high-risk. This means that in the EU, strict obligations must be complied with before such systems can be put on the EU market, including the requirement to undergo a third-party assessment. While broad rules are increasingly being developed, there is often a lack of actionable guidance in how AI can be used in a trustworthy way in a domain as sensitive as justice administration. The United Kingdom and Colombia have sought to put in place such guidance (Box 5.70 and Box 5.71).

Box 5.70. Judicial guidance on AI in the United Kingdom

The UK AI Judicial Guidance provides direction for judicial office holders on the use of AI in courts and tribunals, outlining associated risks and measures to manage them. It ensures that any AI use aligns with judicial obligations, particularly the integrity of justice. The guidance applies to all judicial office holders under the authority of the Lady Chief Justice and Senior President of Tribunals, as well as their clerks and support staff.

The guidance includes key considerations, such as confidentiality, accountability, bias and security. It sets out principles for using AI tools, specifying tasks where AI may assist, such as text summarisation and administrative activities, while advising caution in areas like legal research and analysis. It also highlights the risks of misinformation and the use of AI-generated content in judicial proceedings. In addition to key directions, the document also includes some examples to illustrate potential uses and risks of Generative AI in courts and tribunals.

Developed by a cross-jurisdictional judicial group following consultation with all judicial office holders, the guidance represents an initial step in supporting the judiciary's engagement with AI. Future initiatives include publishing a FAQ document to address questions submitted by the judiciary, with ongoing reviews to adapt to technological advancements.

Source: (United Kingdom, 2023[316])

Box 5.71. Colombia’s directive on AI rules and guidelines for the judiciary

In December 2024, the Superior Council of the Judiciary in Colombia issued Agreement PCSJA24-12243, establishing guidelines for the respectful, responsible, safe, and ethical use of AI within the Judicial Branch. This directive aims to enhance access, transparency, and efficiency in judicial administration by integrating AI into document management, administrative tasks, and judicial support activities. The document outlines three components for implementation: the need for comprehensive training programmes; the implementation of initiatives and sharing of use cases; and the development of guidelines on the use of AI systems in the judiciary.

The directive was preceded by the use of ChatGPT in a ruling concerning an autistic child's exemption from medical fees. In January 2023, a judge in Cartagena incorporated AI-generated responses into his decision to expedite the drafting process. This action sparked significant debate regarding the reliability and ethics of AI in legal judgments. In response, the Superior Council of the Judiciary issued the Agreement PCSJA24-12243, emphasising the need for clear guidelines to ensure AI tools are used appropriately. The incident underscores the urgent need of building digital literacy within the judiciary on AI's capabilities and limitations to prevent over-reliance on technology and to maintain human oversight in judicial decision-making. Since the Agreement PCSJA24-12243, several actions and initiatives have been implemented, including surveys to assess the use of AI by court personnel, training programmes with legal professionals introducing AI-assisted tools for court operations, and partnerships with universities and international organisations on workshops and training programmes.

Source: (Colombia, 2024[317]; Colombia, 2024[317]; UNESCO, 2025[318]).

Untapped potential and way forward

People and legal professionals can use AI applications for services such as legal advice, access to information or document preparation. Well-designed justice systems that incorporate efficiency, transparency and user-friendliness are more likely to be accessible and responsive to the needs of the public. AI’s advanced algorithms and ML techniques can help to ensure that judicial processes are more consistent, transparent and free from human biases, and can help courts to monitor efficiency and resources. AI can improve accountability by flagging inconsistencies and potential biases in real time. Systems can be programmed to monitor for and alert stakeholders to unusual patterns that might indicate bias or misconduct. For example, if data reveals that certain demographic groups are being disproportionately targeted or receiving harsher sentences, these discrepancies can be promptly identified and addressed. Such real-time monitoring and alerts can prompt immediate reviews and corrective actions, enhancing the overall fairness and integrity of the justice system.

The AI systems should be accompanied by pre-existing governance mechanisms on checks and balances (e.g. administrative independence and autonomous judicial review) to ensure the autonomy of the judiciary. Such governance mechanisms align with the vision set forth in the Council of Europe Framework Convention on Artificial Intelligence and Human Rights, Democracy and the Rule of Law (2024[319]), which stresses the importance of preserving judicial independence and access to justice when implementing AI systems.

Trust is perhaps the most crucial aspect in the justice system; decisions can profoundly impact individuals' lives and societal trust. This aspect encompasses transparency, explainability, fairness and accountability, equality of treatment, ethical concerns and privacy and justice gaps that should be carefully addressed. The potential for AI systems to perpetuate or amplify existing societal disparities and justice gaps should be carefully addressed. AI systems should be designed to avoid perpetuating or exacerbating existing biases in the justice system, particularly regarding race, sex or socioeconomic status.

Algorithmic transparency is essential not only for ethical reasons but also to meet legal requirements of due process. The explainability of AI-driven decisions is crucial for maintaining public trust and allowing for meaningful appeals. Ethical considerations should extend to all stakeholders, including defendants, victims, legal professionals and society at large.

AI’s transformative potential in justice requires a human-centred approach that both promotes digital transformation and protects human rights. Effective governance involves collaboration among policymakers, legal professionals and technologists to align AI applications with the principles of fairness, equity and procedural integrity. Continuous impact assessments, stakeholder engagement and mechanisms for redress enhance both legitimacy and effectiveness. This underscores the importance of designing AI systems as tools that support judicial decision-making and justice administration and reinforce the capacity to deliver people-centred justice.

In this context, it remains critical to establish robust governance frameworks, and continuous monitoring mechanisms to ensure the trustworthy and equitable use of AI in justice systems. To support the governance of AI in justice and ensure its responsible use, the OECD AI Principles (Table 2.2), the OECD Framework for Trustworthy AI in Government (Figure 4.6) and the OECD Common Reporting Framework for AI Incidents (2025[320]) can serve as good starting points. The OECD AI Principles advocate for inclusive growth, sustainable development and well-being, emphasising the importance of human-centred values and fairness. They call for AI systems to be transparent, with their decisions understandable by those affected by them. The Framework for Trustworthy AI in Government can be a useful tool to support the responsible use of AI in the justice sector. The OECD Methodology for Monitoring AI Incidents can support the justice sector monitor AI incidents, ensuring these systems are accountable, trustworthy and beneficial, while mitigating potential risks associated with their deployment.

References

[134] Abrell, J., M. Kosch and S. Rausch (2022), “How effective is carbon pricing?—A machine learning approach to policy evaluation”, Journal of Environmental Economics and Management, Vol. 112, p. 102589, https://doi.org/10.1016/j.jeem.2021.102589.

[279] Access Now (2022), Ban Biometric Surveillance, https://www.accessnow.org/campaign/ban-biometric-surveillance/.

[136] ACE (2025), Government administrative data sources for evaluation in Australia, https://evaluation.treasury.gov.au/sites/evaluation.treasury.gov.au/files/2025-01/government-administrative-data-sources-for-evaluation-in-australia.pdf.

[79] Agencia Nacional de Contratación Pública –Colombia Compra Eficiente (2024), VISUALIZACIONES COMPRA PÚBLICA, https://www.colombiacompra.gov.co/content/visualizaciones-compra-publica.

[164] AI.GOV.UK (2024), i.AI Consultation Analyser, https://ai.gov.uk/consultations/.

[223] Aldemir, C. and T. Uçma Uysal (2025), “Artificial Intelligence for Financial Accountability and Governance in the Public Sector: Strategic Opportunities and Challenges”, Administrative Sciences, Vol. 15/2, p. 58, https://doi.org/10.3390/admsci15020058.

[306] Algorithm Watch (2020), Automating Society Report, https://automatingsociety.algorithmwatch.org/report2020/spain/spain-story/.

[130] Alon-Barkat, S. and M. Busuioc (2022), “Human–AI Interactions in Public Sector Decision Making: “Automation Bias” and “Selective Adherence” to Algorithmic Advice”, Journal of Public Administration Research and Theory, Vol. 33/1, pp. 153-169, https://doi.org/10.1093/jopart/muac007.

[216] Alowais, S. et al. (2023), “Revolutionizing healthcare: the role of artificial intelligence in clinical practice”, BMC Medical Education, Vol. 23/1, https://doi.org/10.1186/s12909-023-04698-z.

[270] Amos, J. (2020), “Bushfires: Australian satellite would be ’tuned’ to eucalypt vegetation”, https://www.bbc.com/news/science-environment-51727231 (accessed on 4 March 2025).

[83] Andersson, P., K. Arbin and C. Rosenqvist (2025), “Assessing the value of artificial intelligence (AI) in governmental public procurement”, Journal of Public Procurement, Vol. 25/1, pp. 120-139, https://doi.org/10.1108/jopp-05-2024-0057.

[169] Androutsopoulou, A. et al. (2019), “Transforming the communication between citizens and government through AI-guided chatbots.”, Government information quarterly, Vol. 36/2, pp. 358-367, https://doi.org/10.1016/j.giq.2018.10.001.

[78] Ansari, B., M. Barati and E. Martin (2022), “Enhancing the usability and usefulness of open government data: A comprehensive review of the state of open government data visualization research”, Government Information Quarterly, Vol. 39/1, p. 101657, https://doi.org/10.1016/j.giq.2021.101657.

[193] Anshari, M. et al. (2024), “Public service delivery, artificial intelligence and the sustainable development goals: trends, evidence and complexities”, Journal of Science and Technology Policy Management, Vol. 16/1, pp. 163-181, https://doi.org/10.1108/jstpm-07-2023-0123.

[251] Aoki, N. (2020), “An experimental study of public trust in AI chatbots in the public sector”, Government Information Quarterly, Vol. 37/4, p. 101490, https://doi.org/10.1016/j.giq.2020.101490.

[252] Aoki, N., M. Tay and M. Yarime (2024), “Trustworthy public sector AI: research progress and future agendas”, in Research Handbook on Public Management and Artificial Intelligence, Edward Elgar Publishing, https://doi.org/10.4337/9781802207347.00026.

[55] APA (2023), Worries about artificial intelligence, surveillance at work may be connected to poor mental health, https://www.apa.org/news/press/releases/2023/09/artificial-intelligence-poor-mental-health.

[310] APPR (2025), About the Public Safety Assessment, https://advancingpretrial.org/psa/about/.

[159] Arana-Catania, M. et al. (2021), “Citizen participation and machine learning for a better democracy”, Digital Government: Research and Practice, Vol. 2/3, pp. 1-22, https://doi.org/10.1145/3452118.

[149] Araszkiewicz, M. and V. Rodríguez-Doncel (2019), Legal Knowledge and Information Systems: JURIX 2019: The Thirty-second Annual Conference, https://ebooks.iospress.nl/volume/legal-knowledge-and-information-systems-jurix-2019-the-thirty-second-annual-conference.

[101] Ash, E., S. Galletta and T. Giommoni (2020), “A Machine Learning Approach to Analyzing Corruption in Local Public Finances”, SSRN Electronic Journal, https://doi.org/10.2139/ssrn.3589545.

[77] Attard, J. et al. (2015), “A systematic review of open government data initiatives”, Government Information Quarterly, Vol. 32/4, pp. 399-418, https://doi.org/10.1016/j.giq.2015.07.006.

[10] Australian Government (2020), Department of Veterans’ Affairs Annual Report 2019-20, https://www.transparency.gov.au/publications/veterans-s-affairs/department-of-veterans-affairs/department-of-veterans-affairs-annual-report-2019-20.

[302] Austria Federal Ministry of Justice (2021), Anonymization of court decisions in Austria: Webinar on the use ofAI in the justice field, https://commission.europa.eu/system/files/2021-04/anonymisation_webinar_29032021_austria.pdf (accessed on 19 February 2025).

[198] Babšek, M. et al. (2025), “Artificial Intelligence Adoption in Public Administration: An Overview of Top-Cited Articles and Practical Applications”, AI, Vol. 6/3, p. 44, https://doi.org/10.3390/ai6030044.

[19] Bailey, K. (2023), The Ethics of AI in Finance: How to Detect and Prevent Bias, https://corporatefinanceinstitute.com/resources/data-science/ai-ethics-in-finance-detect-prevent-bias/ (accessed on 10 April 2025).

[217] Bastani, H. et al. (2021), “Efficient and targeted COVID-19 border testing via reinforcement learning”, Nature, Vol. 599/7883, pp. 108-113, https://doi.org/10.1038/s41586-021-04014-z.

[111] Bénassy-Quéré, A. (2022), Ex ante, ex post: tuning the two pillars of policy evaluation, https://www.tresor.economie.gouv.fr/Articles/2022/02/25/ex-ante-ex-post-tuning-the-two-pillars-of-policy-evaluation.

[243] Bennett Institute (2024), Public Health gets personal: The case for an AI-driven personalised prevention platform, https://www.bennettinstitute.cam.ac.uk/wp-content/uploads/2024/03/The-case-for-an-AI-driven-personalised-prevention-platform.pdf.

[165] Berditchevskaia, A. and P. Baeck (2020), The Future of Minds and Machines: How artificial intelligence can enhance collective intelligence.

[35] BEREC (2023), Report on the impact of Artificial Intelligence (AI) solutions in the telecommunications sector on regulation, https://www.berec.europa.eu/en/document-categories/berec/reports/berec-report-on-the-impact-of-artificial-intelligence-ai-solutions-in-the-telecommunications-sector-on-regulation.

[200] Berman, A., K. de Fine Licht and V. Carlsson (2024), “Trustworthy AI in the public sector: An empirical analysis of a Swedish labor market decision-support system”, Technology in Society, Vol. 76, p. 102471, https://doi.org/10.1016/j.techsoc.2024.102471.

[191] Berryhill, J. et al. (2019), “Hello, World: Artificial intelligence and its use in the public sector”, OECD Working Papers on Public Governance, No. 36, OECD Publishing, Paris, https://doi.org/10.1787/726fd39d-en.

[135] Better Evaluation Knowledge (2022), Rapid Evaluation, https://www.betterevaluation.org/methods-approaches/approaches/rapid-evaluation.

[227] Birhane, A. et al. (2022), “The Values Encoded in Machine Learning Research”, 2022 ACM Conference on Fairness, Accountability, and Transparency, https://doi.org/10.1145/3531146.3533083.

[284] Birks, D., M. Townsley and T. Hart (2023), Predictive policing in an Australian context: assessing viability and utility, Australian Institute of Criminology, https://doi.org/10.52922/ti78870.

[114] Blaizot, A. et al. (2022), “Using artificial intelligence methods for systematic review in health sciences: A systematic review”, Research Synthesis Methods, Vol. 13/3, pp. 353-362, https://doi.org/10.1002/jrsm.1553.

[43] Blanc, F. (2018), From Chasing Violations to Managing Risks Origins, challenges and evolutions in regulatory inspections, https://www.elgaronline.com/monobook/9781788112482/9781788112482.xml (accessed on 28 May 2024).

[7] Blanchet, M. and M. Coueffe (2020), Improved GDP nowcasting using large datasets, Ministère de l’Économie et des Finances, Direction générale du Trésor, https://www.tresor.economie.gouv.fr/Articles/2020/04/09/tresor-economics-no-254-improved-gdp-nowcasting-using-large-datasets.

[226] Bloomberg Philanthropies (2023), State of Cities: Generative AI in Local Governments, https://cityaiconnect.jhu.edu/pdfs/Final-Gen-AI-In-Cities-Report_10.18.2023.pdf.

[118] Bohni Nielsen, S., F. Mazzeo Rinaldi and G. Petersson (2024), Artificial Intelligence and Evaluation, Routledge, New York, https://doi.org/10.4324/9781003512493.

[17] Boström, H. et al. (2020), Explaining Multivariate Time Series Forecasts: An Application to Predicting the Swedish GDP, https://ceur-ws.org/Vol-2796/xi-ml-2020_bostrom.pdf.

[189] Bright, J. et al. (2025), “Generative AI is already widespread in the public sector: evidence from a survey of UK public sector professionals”, Digital Government: Research and Practice, Vol. 6/1, pp. 1-13, https://doi.org/10.1145/3700140.

[328] Bright, J. et al. (2019), “Data Science for Local Government”, SSRN Electronic Journal, https://doi.org/10.2139/ssrn.3370217.

[214] Brioscú, A. et al. (2024), “A new dawn for public employment services: Service delivery in the age of artificial intelligence”, OECD Artificial Intelligence Papers, No. 19, OECD Publishing, Paris, https://doi.org/10.1787/5dc3eb8e-en.

[52] Broecke, S. (2023), “Artificial intelligence and labour market matching”, OECD Social, Employment and Migration Working Papers, No. 284, OECD Publishing, Paris, https://doi.org/10.1787/2b440821-en.

[58] Brougham, D. and J. Haar (2017), “Smart Technology, Artificial Intelligence, Robotics, and Algorithms (STARA): Employees’ perceptions of our future workplace”, Journal of Management & Organization, Vol. 24/2, pp. 239-257, https://doi.org/10.1017/jmo.2016.55.

[90] Burger, M., A. Nitsche and J. Arlinghaus (2023), “Hybrid intelligence in procurement: Disillusionment with AI’s superiority?”, Computers in Industry, Vol. 150, https://doi.org/10.1016/j.compind.2023.103946.

[313] Byrom, N., M. Piccinin-Barbieri and P. Wells (2024), “Towards effective governance of justice data”, OECD Working Papers on Public Governance, No. 74, OECD Publishing, Paris, https://doi.org/10.1787/d2950e02-en.

[168] Campagnucci, F. et al. (2025), Artificial Intelligence for Participation, https://scope.uni-muenster.de/wp-content/uploads/2025/01/ENG-Policy-Brief-AI_Jan2025.pdf.

[59] Cappelli, P. and N. Rogovsky (2023), “Artificial intelligence in human resource management”, ILO working paper, https://doi.org/10.54394/ohvv4382.

[187] Carrasco, M. et al. (2019), The Citizen’s Perspective on the Use of AI in Government, https://www.bcg.com/en-ca/publications/2019/citizen-perspective-use-artificial-intelligence-government-digital-benchmarking.

[39] Cary Coglianese (2024), “How to Regulate Artificial Intelligence”, The Regulatory Review, https://www.theregreview.org/2024/01/15/coglianese-how-to-regulate-artificial-intelligence/ (accessed on 28 May 2024).

[291] CEPEJ (2018), European ethical charter on the use of artificial intelligence in judicial systems and their environment, https://rm.coe.int/ethical-charter-en-for-publication-4-december-2018/16808f699c.

[207] Chen, T., M. Gascó-Hernandez and M. Esteve (2023), “The Adoption and Implementation of Artificial Intelligence Chatbots in Public Organizations: Evidence from U.S. State Governments”, The American Review of Public Administration, Vol. 54/3, pp. 255-270, https://doi.org/10.1177/02750740231200522.

[80] ChileCompra (2024), Datos Abiertos de las compras públicas de Chile, https://datos-abiertos.chilecompra.cl/.

[123] Cintron, D. and B. Montrosse-Moorhead (2021), “Integrating Big Data Into Evaluation: R Code for Topic Identification and Modeling”, American Journal of Evaluation, Vol. 43/3, pp. 412-436, https://doi.org/10.1177/10982140211031640.

[297] CMS (2023), CMS Expert Guide to Digital Litigation in Slovenia, https://cms.law/en/int/expert-guides/cms-expert-guide-to-digital-litigation/slovenia (accessed on  June 2024).

[295] CNJ (2025), Labor Court launches artificial intelligence to assist professionals at the institution, https://www.cnj.jus.br/justica-do-trabalho-lanca-inteligencia-artificial-para-auxiliar-profissionais-da-instituicao/.

[204] Cockx, B., M. Lechner and J. Bollens (2023), “Priority to unemployed immigrants? A causal machine learning evaluation of training in Belgium”, Labour Economics, Vol. 80, p. 102306, https://doi.org/10.1016/j.labeco.2022.102306.

[69] Collins, B. (2020), Transformational Results - Purchasing and Strategic Sourcing, https://www2.elpasotexas.gov/municipal-clerk/agenda/04-27-20/1.1.2.pdf.

[317] Colombia (2024), Information System of the Presidential Rapporteurship, https://actosadministrativos.ramajudicial.gov.co/web/Acto%20Administrativo/Default.aspx?ID=19280.

[171] Cortés-Cediel, M. (2023), “Trends and challenges of e-government chatbots: Advances in exploring open government data and citizen participation content.”, Government Information Quarterly, Vol. 40/4, p. 101877, https://doi.org/10.1016/j.giq.2023.101877.

[319] Council of Europe (2024), Framework Convention on Artificial Intelligence and Human Rights, Democracy and the Rule of Law, https://www.coe.int/en/web/artificial-intelligence/the-framework-convention-on-artificial-intelligence.

[293] Court of Cassation of France (2024), The Court of Cassation facing digital technology and artificial intelligence, https://www.vie-publique.fr/parole-dexpert/278415-la-cour-de-cassation-face-au-numerique-et-lintelligence-artificielle.

[18] Crisanto, J. et al. (2024), Regulating AI in the financial sector: recent developments and main challenges, https://www.bis.org/fsi/publ/insights63.pdf.

[280] CSOs (2021), An EU Artificial Intelligence Act for Fundamental Rights: A Civil Society Statement, https://www.accessnow.org/wp-content/uploads/2021/11/joint-statement-EU-AIA.pdf.

[86] da Rosa, I. (2023), Repository of existing projects on transparency in public procurement, Publications Office of the European Union, https://ted.europa.eu/documents/d/ted/final_analysis-report-on-the-inventory-of-contract-registers_-isabel-da-rosa.

[102] de Blasio, G., A. D’Ignazio and M. Letta (2022), “Gotham city. Predicting ‘corrupted’ municipalities with machine learning”, Technological Forecasting and Social Change, Vol. 184, p. 122016, https://doi.org/10.1016/j.techfore.2022.122016.

[308] De Miguel Beriain, I. (2018), “Does the use of risk assessments in sentences respect the right to due process? A critical analysis of the Wisconsin v. Loomis ruling”, Law, Probability and Risk, Vol. 17/1, pp. 45 - 53, https://doi.org/10.1093/lpr/mgy001.

[126] Dell’Acqua, F. et al. (2023), “Navigating the Jagged Technological Frontier: Field Experimental Evidence of the Effects of AI on Knowledge Worker Productivity and Quality”, SSRN Electronic Journal, https://doi.org/10.2139/ssrn.4573321.

[234] Dieterle, E., C. Dede and M. Walker (2024), “The cyclical ethical effects of using artificial intelligence in education”, AI & SOCIETY, Vol. 39/2, pp. 633-643, https://doi.org/10.1007/s00146-022-01497-w.

[48] DRCF (2023), Launch of the International Network for Digital Regulation Cooperation (INDRC), https://www.drcf.org.uk/news-and-events/news/launch-of-international-network-for-digital-regulation-cooperation-indrc.

[66] EC (2020), Study on up-take of emerging technologies, European Commission, https://joinup.ec.europa.eu/sites/default/files/news/2020-06/D.01.06_Final_report_v3.00.pdf.

[218] EC JRC (2021), Selected AI cases in the public sector (JRC129301), http://data.europa.eu/89h/7342ea15-fd4f-4184-9603-98bd87d8239a.

[97] ECA (2024), Special report 05/2024: EU Transparency Register – provides useful but limited information on lobbying activities, https://www.eca.europa.eu/en/publications/SR-2024-05.

[273] ECMWF (2025), ECMWF’s AI forecasts become operational, https://www.ecmwf.int/en/about/media-centre/news/2025/ecmwfs-ai-forecasts-become-operational (accessed on 4 March 2025).

[281] EDPB-EDPS (2021), EDPB-EDPS Joint Opinion 5/2021 on the proposal for a Regulation of the European Parliament and of the Council laying down harmonised rules on artificial intelligence (Artificial Intelligence Act), https://www.edpb.europa.eu/our-work-tools/our-documents/edpbedps-joint-opinion/edpb-edps-joint-opinion-52021-proposal_en.

[127] Eisele, A. et al. (2024), “Exploring the potential of Claude 2 for risk of bias assessment: Using a large language”, p. 10.1101/2024.07.16.24310483.

[172] Eisen, N. et al. (2023), AI can strengthen U.S. democracy—and weaken it, https://www.brookings.edu/articles/ai-can-strengthen-u-s-democracy-and-weaken-it/.

[272] El Morr, C. et al. (2024), “AI-based epidemic and pandemic early warning systems: A systematic scoping review”, Health Informatics Journal, Vol. 30/3, https://doi.org/10.1177/14604582241275844.

[307] Equivant (2025), Solutions: Risk Needs Assessments, https://equivant-supervision.com/solutions/risk-needs-assessments/#COMPAS-R-P.

[174] ERIC (2024), Electronic Registration Information Center (ERIC): Technology and, Electronic Registration Information Center, https://ericstates.org/wp-content/uploads/documents/ERIC-Tech-Security-Brief.pdf.

[285] EUCPN (2022), Artificial intelligence and predictive policing: risks and challenges, https://eucpn.org/document/recommendation-paper-artificial-intelligence-and-predictive-policing-risks-and-challenges.

[155] European Parliament (2022), Report on Artificial Intelligence in a Digital Age, https://www.europarl.europa.eu/doceo/document/A-9-2022-0088_EN.html.

[286] Evans, N. (2022), “Policing the pandemic in Australia and New Zealand: lessons for trust and legitimacy”, Journal of Criminological Research, Policy and Practice, Vol. 9/2, pp. 106-122, https://doi.org/10.1108/jcrpp-10-2022-0050.

[37] Faculty (2021), Automated approaches to measuring online experiences, https://www.ofcom.org.uk/__data/assets/pdf_file/0015/220425/automated-tooling-report.pdf.

[133] Ferretti, S. (2023), “Hacking by the prompt: Innovative ways to utilize ChatGPT for evaluators”, New Directions for Evaluation, Vol. 2023/178-179, pp. 73-84, https://doi.org/10.1002/ev.20557.

[167] Fillet, S. (2024), Cambridge City Council analyzes input 50% faster with Go Vocal’s AI assistant, https://www.govocal.com/case-studies/cambridge-city-council-analyzes-input-50-faster-with-go-vocals-ai-assistant.

[163] Fishkin, J. and A. al (2021), Final Report - LXS 400 - Chile Delibera, https://drive.google.com/file/d/1Vt5XaTk2MW9M_VBN2oztK2nxYZdckl89/view.

[122] Franzen, S. et al. (2022), Advanced Content Analysis: Can Artificial Intelligence Accelerate Theory-Driven Complex Program Evaluation?, World Bank Group, https://documents1.worldbank.org/curated/en/400031645128516191/pdf/Advanced-Content-Analysis-Can-Artificial-Intelligence-Accelerate-Theory-Driven-Complex-Program-Evaluation.pdf.

[13] French Public Finances General Directorate (DGFiP) (2024), Digital Transformation of the Public Finances General Directorate (DGFiP).

[12] French Public Finances General Directorate (DGFiP) (2024), The predictive model.

[62] Friton, P. et al. (2024), Public Procurement of AI: Current and Future Challenges, https://www.lexology.com/library/detail.aspx?g=2c131357-03ed-4965-aa19-e28a65dd9ad6 (accessed on 30 May 2024).

[71] Garcia Rodriguez, M. et al. (2022), “Collusion detection in public procurement auctions with machine learning algorithms”, Automation in Construction, Vol. 133, p. Article 104047, https://doi.org/10.1016/j.autcon.2021.104047.

[64] Gastaldi, L. et al. (2024), AI in Public Settings: Status and Next Steps, https://www.dt.mef.gov.it/export/sites/sitodt/modules/documenti_it/analisi_progammazione/note_tematiche/Nota-Tematica-n-1-February-2024.pdf.

[120] Gatto, L. and P. Bundi (2025), The Use of Quantitative Text Analysis in Evaluations, Routledge, https://doi.org/10.4324/9781003512493-8.

[203] Gesk, T. and M. Leyer (2022), “Artificial intelligence in public services: When and why citizens accept its usage”, Government Information Quarterly, Vol. 39/3, p. 101704, https://doi.org/10.1016/j.giq.2022.101704.

[267] GFDRR (2018), Machine Learning for Disaster Risk Management, http://creativecommons.org/licenses/by/3.0/igo/www.worldbank.org.

[248] Giest, S. and S. Grimmelikhuijsen (eds.) (2020), “Artificial intelligence, bureaucratic form, and discretion in public service”, Information Polity, Vol. 25/4, pp. 491-506, https://doi.org/10.3233/ip-200223.

[65] Glas, A. and F. Kleeman (2016), “The Impact of Industry 4.0 on Procurement and Supply Management: A Conceptual and Qualitative Analysis”, International Journal of Business and Management Invention, Vol. 5/6, pp. 55-66, https://doi.org/10.1590/0103-6513.20180104.

[82] GLOBO (2018), Como as robôs Alice, Sofia e Monica ajudam o TCU a caçar irregularidades em licitações, https://g1.globo.com/economia/tecnologia/noticia/como-as-robos-alice-sofia-e-monica-ajudam-o-tcu-a-cacar-irregularidades-em-licitacoes.ghtml.

[57] Gmyrek, P., J. Berg and D. Bescond (2023), Generative AI and Jobs: A global analysis of potential effects on job quantity and quality, ILO, https://doi.org/10.54394/FHEM8239.

[85] GOV.UK (2019), Guidance - Understanding artificial intelligence ethics and safety, https://www.gov.uk/guidance/understanding-artificial-intelligence-ethics-and-safety.

[180] Government of Iceland (2023), Head Start for Icelandic, https://www.government.is/diplomatic-missions/embassy-article/2023/03/14/Head-start-for-Icelandic/ (accessed on 23 May 2024).

[81] Government of Mexico (2024), Programa Anual de Adquisiciones, Arrendamientos, Servicios y Obra Pública., https://upcp-compranet.hacienda.gob.mx/paaasopdashboard.

[16] Government of Mexico (2023), Plataforma inteligente de apoyos del Gobierno Federal, https://www.transparenciapresupuestaria.gob.mx/es/apoyosdelgobierno (accessed on  20 April 2024).

[300] Government of Peru (2023), Sistema con inteligencia artificial permitirá a jueces resolver solo casos de violencia severo o muy severo, https://www.gob.pe/institucion/pj/noticias/774687-sistema-con-inteligencia-artificial-permitira-a-jueces-resolver-solo-casos-de-violencia-severo-o-muy-severo.

[304] Government of Portugal (2024), Guia Prático da Justiça, https://justica.gov.pt/Servicos/Guia-pratico-da-Justica.

[305] Government of Spain (2007), VioGén, https://www.interior.gob.es/opencms/es/servicios-al-ciudadano/violencia-contra-la-mujer/sistema-viogen/.

[237] Green, B. (2022), “The flaws of policies requiring human oversight of government algorithms”, Computer Law & Security Review, Vol. 45, p. 105681, https://doi.org/10.1016/j.clsr.2022.105681.

[309] Grgić-Hlača, N. et al. (2018), “Beyond Distributive Fairness in Algorithmic Decision Making: Feature Selection for Procedurally Fair Learning”, Proceedings of the AAAI Conference on Artificial Intelligence, Vol. 32/1, https://doi.org/10.1609/aaai.v32i1.11296.

[67] Guida, M. et al. (2023), “The role of artificial intelligence in the procurement process: State of the art and research agenda”, Journal of Purchasing and Supply Management, Vol. 29/2, p. Article 100823, https://doi.org/10.1016/j.pursup.2023.100823.

[27] Guio, A. and K. Müller-Daubermann (2024), Insights and Recommendations for Policy Makers in Latin America and the Caribbean, Global Network of Internet and Society Research Centers, https://networkofcenters.net/sites/networkofcenters.net/files/Summary%20Report%20AI%20Governance%20Costa%20Rica%20NoC.pdf.

[176] Hadfi, R. et al. (2021), “Argumentative Conversational Agents for Online Discussions”, Journal of Systems Science and Systems Engineering, https://doi.org/10.1007/s11518-021-5497-1.

[139] Halpern, D. and D. Maru (2024), Global Evidence Report: A blueprint for better international collaoration on evidence, https://www.bi.team/wp-content/uploads/2024/08/ESRC-Global-Evidence-Report-September-2024.pdf.

[87] Handfield, R., S. Jeong and T. Choi (2019), “Emerging procurement technology: data analytics and cognitive analytics”, International Journal of Physical Distribution & Logistics Management, Vol. 49/10, https://doi.org/10.1108/IJPDLM-11-2017-0348.

[92] Heikkila, M. (2022), AI: Decoded: A Dutch algorithm scandal serves a warning to Europe — The AI Act won’t save us, https://www.politico.eu/newsletter/ai-decoded/a-dutch-algorithm-scandal-serves-a-warning-to-europe-the-ai-act-wont-save-us-2/.

[63] Hickok, M. (2022), Public procurement of artificial intelligence systems: new risks and future proofing., https://doi.org/10.1007/s00146-022-01572-2.

[129] Horowitz, M. (2023), Bending the Automation Bias Curve: A Study of Human and AI-based Decision Making in National Security Contexts, https://arxiv.org/abs/2306.16507.

[260] Hung, T. and C. Yen (2020), “On the person-based predictive policing of AI”, Ethics and Information Technology, Vol. 23/3, pp. 165-176, https://doi.org/10.1007/s10676-020-09539-x.

[257] Hutchins, B. and M. Andrejevic (2021), Olympian Surveillance: Sports Stadiums and the Normalization of Biometric Monitoring, https://ijoc.org/index.php/ijoc/article/view/16377.

[88] IBM (2023), Government Procurement and Acquisition: Opportunities and challenges presented by artificial intelligence and machine learning, IBM Center for The Business of Government, https://www.businessofgovernment.org/sites/default/files/Government%20Procurement%20and%20Acquisition_0.pdf.

[125] Jacob, S. (2025), “Artificial Intelligence and the Future of Evaluation: From Augmented to Automated Evaluation”, Digital Government: Research and Practice, Vol. 6/1, pp. 1-10, https://doi.org/10.1145/3696009.

[202] Jarrahi, M. et al. (2023), “Artificial intelligence and knowledge management: A partnership between human and AI”, Business Horizons, Vol. 66/1, pp. 87-99, https://doi.org/10.1016/j.bushor.2022.03.002.

[221] Jevinger, Å. et al. (2023), “Artificial intelligence for improving public transport: a mapping study”, Public Transport, Vol. 16/1, pp. 99-158, https://doi.org/10.1007/s12469-023-00334-7.

[50] Johnson, B., J. Coggburn and J. Llorens (2022), “Artificial Intelligence and Public Human Resource Management: Questions for Research and Practice”, Public Personnel Management, Vol. 51/4, pp. 538-562, https://doi.org/10.1177/00910260221126498.

[173] Juneja, P. (2024), Artificial Intelligence for Electoral Management, International Institute for Democracy and Electoral Assistance, https://doi.org/10.31752/idea.2024.31.

[5] Jung, J., M. Patnam and A. Ter-Martirosyan (2018), An Algorithmic Crystal Ball: Forecasts-based on Machine Learning, International Monetary Fund, https://www.imf.org/en/Publications/WP/Issues/2018/11/01/An-Algorithmic-Crystal-Ball-Forecasts-based-on-Machine-Learning-46288.

[72] Katona, E. and M. Fazekas (2024), Hidden barriers to open competition: Using text mining to uncover corrupt restrictions to competition in public procurement, https://www.govtransparency.eu/wp-content/uploads/2024/05/Katona-Fazekas_HU_text-miningPP-corruption240417_GTIpublish240430final.pdf.

[250] Kelly, S., S. Kaye and O. Oviedo-Trespalacios (2023), “What factors contribute to the acceptance of artificial intelligence? A systematic review”, Telematics and Informatics, Vol. 77, p. 101925, https://doi.org/10.1016/j.tele.2022.101925.

[215] Khalifa, M. and M. Albadawy (2024), “Artificial Intelligence for Clinical Prediction: Exploring Key Domains and Essential Functions”, Computer Methods and Programs in Biomedicine Update, Vol. 5, p. 100148, https://doi.org/10.1016/j.cmpbup.2024.100148.

[112] Khalil, H., D. Ameen and A. Zarnegar (2022), Tools to support the automation of systematic reviews: a scoping review, J Clin Epidemiology, https://doi.org/10.1016/j.jclinepi.2021.12.005.

[282] Khan, I. et al. (2022), Mandates of the Special Rapporteur (Ref.: OL IRL 3/2022), https://spcommreports.ohchr.org/TMResultsBase/DownLoadPublicCommunicationFile?gId=27594.

[177] Kim, S. et al. (2021), “Moderator Chatbot for Deliberative Discussion”, Proceedings of the ACM on Human-Computer Interaction, Vol. 5/CSCW1, pp. 1-26, https://doi.org/10.1145/3449161.

[20] Klein, A. (2020), Reducing bias in AI-based financial services, https://www.brookings.edu/articles/reducing-bias-in-ai-based-financial-services.

[326] Kleinberg, J. et al. (2017), “Human Decisions and Machine Predictions*”, The Quarterly Journal of Economics, https://doi.org/10.1093/qje/qjx032.

[103] Köbis, N., C. Starke and I. Rahwan (2022), “The promise and perils of using artificial intelligence to fight corruption”, Nature Machine Intelligence, Vol. 4/5, pp. 418-424, https://doi.org/10.1038/s42256-022-00489-1.

[95] Köbis, N., C. Starke and I. Rahwan (2021), Artificial Intelligence as an Anti-Corruption Tool (AI-ACT) -- Potentials and Pitfalls for Top-down and Bottom-up Approaches, https://arxiv.org/abs/2102.11567.

[211] Kopponen, A. et al. (2024), “Personalised public services powered by AI: the citizen digital twin approach”, in Research Handbook on Public Management and Artificial Intelligence, Edward Elgar Publishing, https://doi.org/10.4337/9781802207347.00020.

[9] Korea Fiscal Information Service (KFIS) (2023), Digital Transformation in Public Finance: Korean FMIS (dBrain+).

[241] Labanava, A. et al. (2022), “Capacity Building in Government: Towards Developing a Standard for a Functional Specialist in AI for Public Services”, in Communications in Computer and Information Science, Future Data and Security Engineering. Big Data, Security and Privacy, Smart City and Industry 4.0 Applications, Springer Nature Singapore, Singapore, https://doi.org/10.1007/978-981-19-8069-5_34.

[116] Lai, H. et al. (2024), “Assessing the Risk of Bias in Randomized Clinical Trials With Large Language Models”, JAMA Network Open, Vol. 7/5, p. e2412687, https://doi.org/10.1001/jamanetworkopen.2024.12687.

[156] Landemore, H. (2022), “Can AI bring deliberative democracy to the masses?”, https://www.law.nyu.edu/sites/default/files/Helen%20Landemore%20Can%20AI%20bring%20deliberative%20democracy%20to%20the%20masses.pdf.

[283] Lecorps, Y. and G. Tissandier (2022), “PAVED With Good Intentions? An Evaluation of a French Police Predictive Policing System”, SSRN Electronic Journal, https://doi.org/10.2139/ssrn.4314831.

[121] Leeuw, F. (2003), “Reconstructing Program Theories: Methods Available and Problems to be Solved”, The American Journal of Evaluation, Vol. 24/1, pp. 5-20, https://doi.org/10.1016/s1098-2140(02)00271-0.

[196] Leyer, M. and S. Schneider (2021), “Decision augmentation and automation with artificial intelligence: Threat or opportunity for managers?”, Business Horizons, Vol. 64/5, pp. 711-724, https://doi.org/10.1016/j.bushor.2021.02.026.

[321] Linders, D., C. Liao and C. Wang (2018), “Proactive e-Governance: Flipping the service delivery model from pull to push in Taiwan”, Government Information Quarterly, Vol. 35/4, pp. S68-S76, https://doi.org/10.1016/j.giq.2015.08.004.

[73] López-Iturriaga, F. and I. Sanz (2017), “Predicting Public Corruption with Neural Networks: An Analysis of Spanish Provinces”, Social Indicators Research, Vol. 140/3, pp. 975-998, https://doi.org/10.1007/s11205-017-1802-2.

[190] Lorenz, P., K. Perset and J. Berryhill (2023), “Initial policy considerations for generative artificial intelligence”, OECD Artificial Intelligence Papers, No. 1, OECD Publishing, Paris, https://doi.org/10.1787/fae2d1e6-en.

[220] Macnish, K., D. Wright and T. Jiya (2021), “Correction to: Predictive Policing in 2025: A Scenario”, in Advanced Sciences and Technologies for Security Applications, Policing in the Era of AI and Smart Societies, Springer International Publishing, Cham, https://doi.org/10.1007/978-3-030-50613-1_13.

[194] Madan, R. and M. Ashok (2023), “AI adoption and diffusion in public administration: A systematic literature review and future research agenda”, Government Information Quarterly, Vol. 40/1, p. 101774, https://doi.org/10.1016/j.giq.2022.101774.

[128] Marcucci, S. and S. Verhulst (2025), Reimagining the Policy Cycle in the Age of Artificial Intelligence, Elsevier BV, https://doi.org/10.2139/ssrn.5137557.

[185] Margetts, H. (2011), “The Internet and Transparency”, The Political Quarterly, Vol. 82/4, pp. 518-521, https://doi.org/10.1111/j.1467-923x.2011.02253.x.

[322] Margetts, H. and C. Dorobantu (2019), “Rethink government with AI”, Nature, Vol. 568/7751, pp. 163-165, https://doi.org/10.1038/d41586-019-01099-5.

[33] Martin, P. (2023), Policy modelling with Rules as Code (RaC), https://salsa.digital/insights/policy-modelling-with-rules-as-code-rac.

[261] McKendrick, K. (2019), Artificial Intelligence Prediction and Counterterrorism, https://www.chathamhouse.org/sites/default/files/2019-08-07-AICounterterrorism.pdf.

[238] Medaglia, R., J. Gil-Garcia and T. Pardo (2021), “Artificial Intelligence in Government: Taking Stock and Moving Forward”, Social Science Computer Review, Vol. 41/1, pp. 123-140, https://doi.org/10.1177/08944393211034087.

[239] Medaglia, R., P. Mikalef and L. Tangi (2025), Competences and governance practices for artificial intelligence in the public sector, European Commission: Joint Research Centre, https://doi.org/10.2760/7895569.

[53] Milanez, A., A. Lemmens and C. Ruggiu (2025), “Algorithmic management in the workplace: New evidence from an OECD employer survey”, OECD Artificial Intelligence Papers, No. 31, OECD Publishing, Paris, https://doi.org/10.1787/287c13c4-en.

[124] Miller, S. (2020), “Causal Forest estimation of heterogeneous and time-varying environmental policy effects”, journal of environmental Economics and Management, Vol. 103, https://doi.org/10.1016/j.jeem.2020.102337.

[195] Miller, S. and L. Keiser (2020), “Representative Bureaucracy and Attitudes Toward Automated Decision Making”, Journal of Public Administration Research and Theory, Vol. 31/1, pp. 150-165, https://doi.org/10.1093/jopart/muaa019.

[115] Minozzi, S. et al. (2020), The revised Cochrane risk of bias tool for randomized trials (RoB 2) showed low interrater reliability and challenges in its application, Journal of Clinical Epidemiology, https://doi.org/10.1016/j.jclinepi.2020.06.015.

[148] MITRE (2023), Searching for Solutions: MITRE Tool Simplifies Freedom of Information Act Requests, https://www.mitre.org/news-insights/impact-story/mitre-tool-simplifies-freedom-information-act-requests.

[4] Mohun, J. and A. Roberts (2020), “Cracking the code: Rulemaking for humans and machines”, OECD Working Papers on Public Governance, No. 42, OECD Publishing, Paris, https://doi.org/10.1787/3afe6ba5-en.

[104] Mpofu, F. (2023), “The application of Artificial Intelligence in external auditing and its implications on audit quality? A review of the ongoing debates”, International Journal of Research in Business and Social Science (2147- 4478), Vol. 12/9, pp. 496-512, https://doi.org/10.20525/ijrbs.v12i9.2737.

[119] Næss, T. et al. (2025), Text Mining and Machine Learning in a Performance Audit of Police Handling of Cybercrime in Norway, Routledge, https://doi.org/10.4324/9781003512493.

[150] National Documentation Center (EKT) (2025), , https://www.didaktorika.gr/eadd/.

[70] NCDIT (2024), New Chatbot Helps Answer IT Procurement Questions, https://it.nc.gov/blog/2024/04/04/new-chatbot-helps-answer-it-procurement-questions (accessed on 19 February 2024).

[269] Nelson, D., Y. He and G. Moore (2024), “Trends and applications in wildfire burned area mapping: Remote sensing data, cloud geoprocessing platforms, and emerging algorithms”, Geomatica, Vol. 76/1, p. 100008, https://doi.org/10.1016/j.geomat.2024.100008.

[32] Nesta (2024), Innovation Policy Simulation for the Smart Economy, https://www.nesta.org.uk/feature/smarter-policy-through-simulation/innovation-policy-simulation-for-the-smart-economy/.

[138] Netherlands Ministry of the Interior and Kingdom Relations (2019), NL Digital Data Agenda Government, https://www.nldigitalgovernment.nl/wp-content/uploads/sites/11/2019/04/data-agenda-government.pdf.

[230] Nikiforova, A. et al. (2023), PPPS’2023-Proactive and Personalised Public Services: Searching for Meaningful Human Control in Algorithmic Government, https://ceur-ws.org/Vol-3449/paper27.pdf.

[117] Odell, C. (2024), Why we need to explore the use of AI in evidence synthesis: Reflections from the Global Evidence Summit 2024, https://researchforevidence.fhi360.org/need-explore-use-ai-evidence-synthesis-reflections-global-evidence-summit-2024.

[99] Odilla, F. (2023), “Bots against corruption: Exploring the benefits and limitations of AI-based anti-corruption technology”, Crime, Law and Social Change, Vol. 80/4, pp. 353-396, https://doi.org/10.1007/s10611-023-10091-0.

[212] OECD (2025), “AI and the future of social protection in OECD countries”, OECD Artificial Intelligence Papers, No. 42, OECD Publishing, Paris, https://doi.org/10.1787/7b245f7e-en.

[192] OECD (2025), How Innovation Ecosystems Foster Citizen Participation Using Emerging Technologies in Portugal, Spain and the Netherlands, OECD Public Governance Reviews, OECD Publishing, Paris, https://doi.org/10.1787/2cb37a30-en.

[107] OECD (2025), Implementation Toolkit for the OECD Recommendation on Public Policy Evaluation, OECD Public Governance Reviews, OECD Publishing, Paris, https://doi.org/10.1787/77faa4fe-en.

[31] OECD (2025), OECD Regulatory Policy Outlook 2025, OECD Publishing, Paris, https://doi.org/10.1787/56b60e39-en.

[266] OECD (2025), Prevention of Stalking Crimes Powered by Smart CCTV Video Analysis Technology, https://oecd-opsi.org/innovations/prevention-of-stalking-crimes-powered-by-smart-cctv-video-analysis-technology/ (accessed on 12 March 2025).

[183] OECD (2025), “Tackling civic participation challenges with emerging technologies: Beyond the hype”, OECD Public Governance Policy Papers, No. 72, OECD Publishing, Paris, https://doi.org/10.1787/ec2ca9a2-en.

[320] OECD (2025), “Towards a common reporting framework for AI incidents”, OECD Artificial Intelligence Papers, No. 34, OECD Publishing, Paris, https://doi.org/10.1787/f326d4ac-en.

[188] OECD (2024), “2023 OECD Digital Government Index: Results and key findings”, OECD Public Governance Policy Papers, No. 44, OECD Publishing, Paris, https://doi.org/10.1787/1a89ed5e-en.

[244] OECD (2024), AI in Health: Huge potential, huge risks, https://www.oecd.org/health/AI-in-health-huge-potential-huge-risks.pdf.

[93] OECD (2024), Anti-Corruption and Integrity Outlook 2024, OECD Publishing, Paris, https://doi.org/10.1787/968587cd-en.

[29] OECD (2024), “Assessing potential future artificial intelligence risks, benefits and policy imperatives”, OECD Artificial Intelligence Papers, No. 27, OECD Publishing, Paris, https://doi.org/10.1787/3f4e3dfb-en.

[146] OECD (2024), Enabling Digital Innovation in Government: The OECD GovTech Policy Framework, OECD Digital Government Studies, OECD Publishing, Paris, https://doi.org/10.1787/a51eb9b2-en.

[24] OECD (2024), Facts not Fakes: Tackling Disinformation, Strengthening Information Integrity, OECD Publishing, Paris, https://doi.org/10.1787/d909ff7a-en.

[21] OECD (2024), “Financial Management Information Systems in OECD countries”, OECD Papers on Budgeting, No. 2024/02, OECD Publishing, Paris, https://doi.org/10.1787/ce8367cd-en.

[253] OECD (2024), “Fixing frictions: ‘sludge audits’ around the world: How governments are using behavioural science to reduce psychological burdens in public services”, OECD Public Governance Policy Papers, No. 48, OECD Publishing, Paris, https://doi.org/10.1787/5e9bb35c-en.

[144] OECD (2024), Framework for the Anticipatory Governance of Emerging Technologies, https://doi.org/10.1787/0248ead5-en.

[201] OECD (2024), Global Trends in Government Innovation 2024: Fostering Human-Centred Public Services, OECD Public Governance Reviews, OECD Publishing, Paris, https://doi.org/10.1787/c1bc19c3-en.

[60] OECD (2024), “Governing with Artificial Intelligence: Are governments ready?”, OECD Artificial Intelligence Papers, No. 20, OECD Publishing, Paris, https://doi.org/10.1787/26324bc2-en.

[210] OECD (2024), Modernising Access to Social Protection: Strategies, Technologies and Data Advances in OECD Countries, OECD Publishing, Paris, https://doi.org/10.1787/af31746d-en.

[259] OECD (2024), OECD Digital Economy Outlook 2024 (Volume 1): Embracing the Technology Frontier, OECD Publishing, Paris, https://doi.org/10.1787/a1689dc5-en.

[46] OECD (2024), OECD Survey on Drivers of Trust in Public Institutions – 2024 Results: Building Trust in a Complex Policy Environment, OECD Publishing, Paris, https://doi.org/10.1787/9a20554b-en.

[61] OECD (2024), Public procurement, https://www.oecd.org/en/topics/public-procurement.html.

[91] OECD (2024), Recommendation of the Council on Artificial Intelligence, https://legalinstruments.oecd.org/en/instruments/OECD-LEGAL-0449 (accessed on 2 June 2024).

[142] OECD (2024), Reinforcing democracy initiative, https://www.oecd.org/en/about/programmes/reinforcing-democracy-initiative.

[224] OECD (2024), Shaping smart cities of all sizes, OECD Publishing, https://oe.cd/cities-2024-roundtable.

[75] OECD (2024), Strengthening Oversight of the Court of Auditors for Effective Public Procurement in Portugal: Digital Transformation and Data-driven Risk Assessments, OECD Public Governance Reviews, OECD Publishing, Paris, https://doi.org/10.1787/35aeab1e-en.

[315] OECD (2023), 2023 OECD Open, Useful and Re-usable data (OURdata) Index: Results and key findings, OECD Publishing, https://doi.org/10.1787/a37f51c3-en.

[182] OECD (2023), “AI language models: Technological, socio-economic and policy considerations”, OECD Digital Economy Papers, No. 352, OECD Publishing, Paris, https://doi.org/10.1787/13d38f92-en.

[231] OECD (2023), “Emerging privacy-enhancing technologies: Current regulatory and policy approaches”, OECD Digital Economy Papers, No. 351, OECD Publishing, Paris, https://doi.org/10.1787/bf121be4-en.

[141] OECD (2023), Engaging citizens in innovation policy: Why, when and how?, OECD Publishing, https://doi.org/10.1787/ba068fa6-en.

[2] OECD (2023), Global Trends in Government Innovation 2023, OECD Public Governance Reviews, OECD Publishing, Paris, https://doi.org/10.1787/0655b570-en.

[34] OECD (2023), Government at a Glance 2023, OECD Publishing, Paris, https://doi.org/10.1787/3d5c5d31-en.

[54] OECD (2023), OECD Employment Outlook 2023: Artificial Intelligence and the Labour Market, OECD Publishing, Paris, https://doi.org/10.1787/08785bba-en.

[151] OECD (2023), Open Government for Stronger Democracies: A Global Assessment, OECD Publishing, https://doi.org/10.1787/5478db5b-en.

[51] OECD (2023), Public Employment and Management 2023: Towards a Flexible Public Service, OECD Publishing, Paris, https://doi.org/10.1787/5b378e11-en.

[289] OECD (2023), Recommendation of the Council on Access to Justice and People-Centred Justice Systems, https://legalinstruments.oecd.org/en/instruments/OECD-LEGAL-0498.

[225] OECD (2023), Smart City Data Governance: Challenges and the Way Forward, OECD Urban Studies, OECD Publishing, Paris, https://doi.org/10.1787/e57ce301-en.

[23] OECD (2022), “2022 OECD Financial Management and Reporting Survey”, https://data-explorer.oecd.org (accessed on 22 April 2024).

[143] OECD (2022), Building Trust to Reinforce Democracy: Main Findings from the 2021 OECD Survey on Drivers of Trust in Public Institutions, Building Trust in Public Institutions, OECD Publishing, Paris, https://doi.org/10.1787/b407f99c-en.

[256] OECD (2022), Exploring Innovation in Law Enforcement: Opportunities and critical considerations, OECD Publishing, https://oecd-opsi.org/blog/exploring-innovation-in-law-enforcement/.

[184] OECD (2022), OECD Guidelines for Citizen Participation Processes, OECD Publishing, https://doi.org/10.1787/f765caf6-en.

[246] OECD (2022), Supporting Health Innovation with Fair Information Practice Principles, https://www.oecd.org/content/dam/oecd/en/topics/policy-sub-issues/digital-health-systems/oecd-israel-health-data-governance-workshop-report.pdf.

[1] OECD (2022), Tax Administration 2022: Comparative Information on OECD and other Advanced and Emerging Economies, OECD Publishing, Paris, https://doi.org/10.1787/1e797131-en.

[152] OECD (2022), The Protection and Promotion of Civic Space: Strengthening Alignment with International Standards and Guidance, OECD Publishing, https://doi.org/10.1787/d234e975-en.

[96] OECD (2021), Countering Public Grant Fraud in Spain: Machine Learning for Assessing Risks and Targeting Control Activities, OECD Public Governance Reviews, OECD Publishing, Paris, https://doi.org/10.1787/0ea22484-en.

[25] OECD (2021), Data-Driven, Information-Enabled Regulatory Delivery, OECD Publishing, Paris, https://doi.org/10.1787/8f99ec8c-en.

[175] OECD (2021), Evaluation Guidelines for Representative Deliberative Processes, OECD Publishing, https://doi.org/10.1787/10ccbfcb-en.

[28] OECD (2021), OECD Regulatory Policy Outlook 2021, OECD Publishing, Paris, https://doi.org/10.1787/38b0fdb1-en.

[49] OECD (2021), Public Employment and Management 2021: The Future of the Public Service, OECD Publishing, Paris, https://doi.org/10.1787/938f0d65-en.

[329] OECD (2021), Recommendation of the Council for Agile Regulatory Governance to Harness Innovation, https://legalinstruments.oecd.org/en/instruments/OECD-LEGAL-0464.

[47] OECD (2021), Recommendation of the Council on Enhancing Access to and Sharing of Data, OECD Publishing, https://legalinstruments.oecd.org/en/instruments/OECD-LEGAL-0463.

[254] OECD (2020), Artificial Intelligence & Responsible Business Conduct, https://mneguidelines.oecd.org/RBC-and-artificial-intelligence.pdf.

[108] OECD (2020), Improving Governance with Policy Evaluation: Lessons From Country Experiences, OECD Public Governance Reviews, OECD Publishing, Paris, https://doi.org/10.1787/89b1577d-en.

[161] OECD (2020), Innovative Citizen Participation and New Democratic Institutions: Catching the Deliberative Wave, OECD Publishing, https://doi.org/10.1787/339306da-en.

[255] OECD (2020), Public Provider versus Big Brother, OECD Publishing, https://trends.oecd-opsi.org/trend-reports/public-provider-versus-big-brother/.

[36] OECD (2020), Shaping the Future of Regulators: The Impact of Emerging Technologies on Economic Regulators, The Governance of Regulators, OECD Publishing, Paris, https://doi.org/10.1787/db481aa3-en.

[3] OECD (2020), Tax Administration 3.0: The Digital Transformation of Tax Administration, OECD Publishing, Paris, https://doi.org/10.1787/ca274cc5-en.

[278] OECD (2020), “Tracking and tracing COVID: Protecting privacy and data while using apps and biometrics”, OECD Policy Responses to Coronavirus (COVID-19), OECD Publishing, Paris, https://doi.org/10.1787/8f394636-en.

[42] OECD (2018), OECD Regulatory Enforcement and Inspections Toolkit, OECD Publishing, Paris, https://doi.org/10.1787/9789264303959-en.

[245] OECD (2017), Recommendation of the Council on Health Data Governance, https://legalinstruments.oecd.org/en/instruments/OECD-LEGAL-0433.

[140] OECD (2017), Recommendation of the Council on Open Government, https://legalinstruments.oecd.org/en/instruments/OECD-LEGAL-0438.

[74] OECD (2015), Recommendation of the Council on Public Procurement, https://legalinstruments.oecd.org/en/instruments/OECD-LEGAL-0411 (accessed on 2 June 2024).

[41] OECD (2014), Regulatory Enforcement and Inspections, OECD Best Practice Principles for Regulatory Policy, OECD Publishing, Paris, https://doi.org/10.1787/9789264208117-en.

[153] OECD OPSI (2019), Kol Zchut, https://oecd-opsi.org/innovations/kol-zchut/.

[298] OECD/CAF (2022), The Strategic and Responsible Use of Artificial Intelligence in the Public Sector of Latin America and the Caribbean, OECD Public Governance Reviews, OECD Publishing, Paris, https://doi.org/10.1787/1f334543-en.

[40] OECD/EU (2024), Reforming Regulatory Inspections in Italy at Regional and National Level 21IT12 (Final Report), https://reform-support.ec.europa.eu/document/download/e2ad74a6-9bf9-4eef-8ffc-a8f2aaee3ccd_cs?filename=05nov2024%20Final%20Report%2021IT12%20Reforming%20inspections%20in%20Italy.pdf.

[131] OECD/UNESCO (2024), G7 Toolkit for Artificial Intelligence in the Public Sector, OECD Publishing, Paris, https://doi.org/10.1787/421c1244-en.

[327] Ojo, A., S. Mellouli and F. Ahmadi Zeleti (2019), “A Realist Perspective on AI-era Public Management*”, Proceedings of the 20th Annual International Conference on Digital Government Research, https://doi.org/10.1145/3325112.3325261.

[158] Ovadya, A. (2023), “Bridging Systems: Open Problems for Countering Destructive Divisiveness across Ranking, Recommenders, and Governance”, Arxiv, https://arxiv.org/abs/2301.09976.

[14] Palkeet (2024), RPA/AI in Financial Management.

[222] Pawlowski, C. and H. Scholta (2023), “A taxonomy for proactive public services”, Government Information Quarterly, Vol. 40/1, p. 101780, https://doi.org/10.1016/j.giq.2022.101780.

[178] Peixoto, T., O. Canuto and L. Jordan (2024), AI and the Future of Government: Unexpected Effects and Critical Challenges, https://www.policycenter.ma/publications/ai-and-future-government-unexpected-effects-and-critical-challenges.

[233] Perry, A. and N. Lee (2019), AI is coming to schools, and if we’re not careful, so will its biases, https://www.brookings.edu/articles/ai-is-coming-to-schools-and-if-were-not-careful-so-will-its-biases/.

[287] Perry, W. et al. (2013), Predictive Policing: The Role of Crime Forecasting in Law Enforcement Operations, RAND Corporation, Santa Monica, CA, https://doi.org/10.7249/RR233.

[276] PetaBencana.id (2021), As Jakarta floods again, humanitarian chatbots on social media support community-led disaster response, https://info.petabencana.id/2021/02/22/as-jakarta-floods-again-humanitarian-chatbots-on-social-media-support-community-led-disaster-response/ (accessed on 4 March 2025).

[109] Petersson, G. et al. (2017), Cyber Society, Big Data, and Evaluation, Routledge, https://www.routledge.com/Cyber-Society-Big-Data-and-Evaluation/Petersson-Breul/p/book/9781138483033?srsltid=AfmBOooqz5iNGr7Tpm5k4aiNFcJc0RAx_yZwMW42X6AiuvNZIs0Q3bi_.

[288] Petra, B. et al. (2024), Strategies to Counter Artificial Intelligence in Law Enforcement: Cross-Country Comparison of Citizens in Greece, Italy and Spain, https://arxiv.org/abs/2405.19970.

[213] Policy Lab Digital, Work & Society within the German Federal Ministry of Labour and Social Affairs (2024), Guidelines for the Use of AI in the Administrative Work of Employment and Social Protection Services, https://www.denkfabrik-bmas.de/fileadmin/Downloads/Publikationen/Guidelines_for_the_use_of_ai_in_the_administrative_work_of_employment_and_social_protection_services.pdf.

[45] Productivity Committee (2024), Making the Most of the AI Opportunity: Data Availability and Use, https://www.pc.gov.au/research/completed/making-the-most-of-the-ai-opportunity/ai-paper3-data.pdf.

[311] Queen Mary University of London; Centre for European Policy Studies (2021), Criminal Justice, Fundamental Rights and the Rule of Law in the Digital Age, https://cdn.ceps.eu/wp-content/uploads/2021/05/Criminal-Justice-Fundamental-Rights-and-the-Rule-of-law-in-the-Digital-Age.pdf.

[206] Ramires Hernández, P., D. Valle-Cruz and R. Mendoza Méndez (2022), “Review on the Application of Artificial Intelligence-Based Chatbots in Public Administration”, in Handbook of Research on Applied Artificial Intelligence and Robotics for Government Processes, Advances in Computational Intelligence and Robotics, IGI Global, https://doi.org/10.4018/978-1-6684-5624-8.ch007.

[249] Rane, N., S. Choudhary and J. Rane (2024), “Artificial intelligence for enhancing resilience”, Journal of Applied Artificial Intelligence, Vol. 5/2, pp. 1-33, https://doi.org/10.48185/jaai.v5i2.1053.

[229] Redden, J. et al. (2022), Automating Public Services: Learning from Cancelled Systems, Carnegie UK, https://carnegieuk.org/publication/automating-public-services-learning-from-cancelled-systems/.

[290] Redden, J. and M. O’Donovan Dix (2020), Artificial intelligence in the criminal justice system: Demystifying artificial intelligence, its applications, and potential risks, https://www.ojp.gov/ncjrs/virtual-library/abstracts/artificial-intelligence-criminal-justice-system-demystifying.

[292] Reiling, D. (2020), Courts and artificial intelligence, https://doi.org/10.36745/ijca.343.

[6] Richardson, A., T. van Florenstein Mulder and T. Vehbi (2019), Nowcasting GDP using machine learning algorithms: A real-time assessment, https://www.rbnz.govt.nz/-/media/c0bb0238351244c986e33550b36bd2d4.ashx.

[110] Rinaldi, F., G. Giuffrida and T. Negrete (2017), Real-Time Monitoring and Evaluation—Emerging News as Predictive Process Using Big Data-Based Approach, Routledge.

[22] Rivero del Paso, L. et al. (2023), Digital Solutions Guidelines for Public Financial Management, https://www.imf.org/en/Publications/TNM/Issues/2023/10/06/Digital-Solutions-Guidelines-for-Public-Financial-Management-537781.

[179] Romberg, J. and T. Escher (2024), “Making Sense of Citizens’ Input through Artificial Intelligence: A Review of Methods for Computational Text Analysis to Support the Evaluation of Contributions in Public Participatio”, Digital Government: Research and Practice, pp. 1-30, https://doi.org/10.1145/3603254.

[100] Salazar, A., J. Pérez and J. Gallego (2024), “VigIA: prioritizing public procurement oversight with machine learning models and risk indices”, Data & Policy, Vol. 6, https://doi.org/10.1017/dap.2024.83.

[76] Santiso, C. (2022), “Govtech against corruption: What are the integrity dividends of government digitalization?”, Data & Policy, Vol. 4, https://doi.org/10.1017/dap.2022.31.

[166] Schneider, B. and N. Sanders (2023), AI could shore up democracy - here’s one way, https://theconversation.com/ai-could-shore-up-democracy-heres-one-way-207278 (accessed on 23 May 2024).

[205] SDAIA (2024), SDAIA Showcases National Smart C Platform and National Algorithms for Smart Cities at Barcelona Expo, https://sdaia.gov.sa/en/MediaCenter/News/Pages/NewsDetails.aspx?NewsID=228.

[271] Seto, J. (2018), Using AI and IoT for disaster management, https://azure.microsoft.com/en-us/blog/using-ai-and-iot-for-disaster-management/ (accessed on 3 March 2025).

[236] Shark, A. (2025), What the Rising Costs of AI Means for Government, https://statetechmagazine.com/article/2025/01/what-rising-costs-ai-means-government.

[84] Shark, A. (2024), Artificial Intelligence: The Impact on Public Procurement - Opportunity & Risk, https://www.nigp.org/blog/ai-and-public-procurement.

[296] Slovenia Ministry of Justice (2024), Information request to Maja Velic, advisor of the Ministry of Justice of Slovenia.

[228] Soman, D. (2024), What Works, What Doesn’t (and When): Case Studies in Applied Behavioral Science, https://utppublishing.com/doi/book/10.3138/9781487548735.

[303] Sommer, J. (2021), La Cour de cassation à l’épreuve du numérique et de l’intelligence artificielle, https://www.vie-publique.fr/parole-dexpert/278415-la-cour-de-cassation-face-au-numerique-et-lintelligence-artificielle (accessed on 19 February 2025).

[186] Song, C. and J. Lee (2015), “Citizens’ Use of Social Media in Government, Perceived Transparency, and Trust in Government”, Public Performance & Management Review, Vol. 39/2, pp. 430-453, https://doi.org/10.1080/15309576.2015.1108798.

[314] Spain Ministry of Justice (2024), Política de uso de la IA en la Administración de Justicia, https://www.administraciondejusticia.gob.es/w/aprobada-la-politica-de-uso-de-la-ia-en-la-administracion-de-justicia.

[301] Spain Ministry of Justice (2022), Digital transformation within the adminisitration of justice, https://www.mjusticia.gob.es/es/JusticiaEspana/ProyectosTransformacionJusticia/Documents/202209%20Digital%20Transformation%20[ENG].pdf (accessed on 19 February 2025).

[137] Statistics Denmark (2025), Microdata schemes in Statistics Denmark, https://www.dst.dk/en/TilSalg/Forskningsservice/Dataadgang/Projektoprettelse.

[277] Stettinger, G., P. Weissensteiner and S. Khastgir (2024), “Trustworthiness Assurance Assessment for High-Risk AI-Based Systems”, IEEE Access, Vol. 12, pp. 22718-22745, https://doi.org/10.1109/access.2024.3364387.

[299] STF (2024), STF lança MARIA, ferramenta de inteligência artificial que dará mais agilidade aos serviços do Tribunal, https://noticias.stf.jus.br/postsnoticias/stf-lanca-maria-ferramenta-de-inteligencia-artificial-que-dara-mais-agilidade-aos-servicos-do-tribunal/.

[199] Straub, V. et al. (2024), AI for bureaucratic productivity: Measuring the potential of AI to help automate 143 million UK government transactions, https://www.turing.ac.uk/sites/default/files/2024-03/ai_for_bureaucratic_productivity.pdf.

[242] Sumner, J. et al. (2023), “Developing an Artificial Intelligence-Driven Nudge Intervention to Improve Medication Adherence: A Human-Centred Design Approach”, Journal of Medical Systems, Vol. 48/1, https://doi.org/10.1007/s10916-023-02024-0.

[38] Sunass (2024), Use of big data and AI in water regulation: Experience from Sunass, Peru, https://cdn.www.gob.pe/uploads/document/file/6872996/5943561-presentacion-en-la-red-de-reguladores-economicos-o-network-economic-regulators-ner.pdf.

[197] Sun, T. and R. Medaglia (2019), “Mapping the challenges of Artificial Intelligence in the public sector: Evidence from public healthcare”, Government Information Quarterly, Vol. 36/2, pp. 368-383, https://doi.org/10.1016/j.giq.2018.09.008.

[294] Supreme Court of Brazil (2024), Information request to Victor Durigan, advisor to the Supreme Court.

[94] Taşdöken, Ö. (2024), “Use of Artificial Intelligence and Audit Analytics in Internal Audit Processes in the Public Sector”, EDPACS, Vol. 69/9, pp. 1-15, https://doi.org/10.1080/07366981.2024.2376790.

[219] TBI (2024), The Economic Case for Reimagining the State, Tony Blair Institute for Global Change, https://institute.global/insights/economic-prosperity/the-economic-case-for-reimagining-the-state.

[209] Tejwani, R. et al. (2020), Migratable AI: Effect of identity and information migration on users perception of conversational AI agents, https://ieeexplore.ieee.org/document/9223436.

[8] Tenorio, J. and W. Perez (2023), GDP nowcasting with Machine Learning and Unstructured Data to Peru, https://perueconomics.org/wp-content/uploads/2023/11/WP-197.pdf.

[132] The Alan Turing Institute (2023), AI Skills for Business Competency Framework, https://www.turing.ac.uk/skills/collaborate/ai-skills-business-framework.

[162] The Computational Democracy Project (2024), The Computational Democracy Project, https://compdemocracy.org/.

[68] The World Bank (2023), OUTPUT 14 - Report on the proposed tool for screening of legislation, https://documents1.worldbank.org/curated/en/099011924055735846/pdf/P169141129e0ad0f618892137a09a4eb9f9.pdf.

[235] Thiebes, S., S. Lins and A. Sunyaev (2020), “Trustworthy artificial intelligence”, Electronic Markets, Vol. 31/2, pp. 447-464, https://doi.org/10.1007/s12525-020-00441-4.

[113] Thomas, J. et al. (2024), Responsible AIin Evidence SynthEsis (RAISE): guidance and recommendations, https://osf.io/fwaud/.

[268] Tilon, S. et al. (2020), “Post-Disaster Building Damage Detection from Earth Observation Imagery Using Unsupervised and Transferable Anomaly Detecting Generative Adversarial Networks”, Remote Sensing, Vol. 12/24, p. 4193, https://doi.org/10.3390/rs12244193.

[240] Trajkovski, G. (2024), “Bridging the public administration‐AI divide: A skills perspective”, Public Administration and Development, Vol. 44/5, pp. 412-426, https://doi.org/10.1002/pad.2061.

[26] Tribune News Service (2023), State Legislature Adopts Resolution on AI ... Drafted by AI, https://insider.govtech.com/california/news/state-legislature-adopts-resolution-on-ai-drafted-by-ai.

[160] Tsai, L. et al. (2024), “Generative AI for Pro-Democracy Platforms”, An MIT Exploration of Generative AI, https://doi.org/10.21428/e4baedd9.5aaf489a.

[106] U4 (2025), Unlocking AI’s potential in anti-corruption: Hype vs. reality, https://www.u4.no/blog/unlocking-ai-s-potential-in-anti-corruption-hype-vs-reality.

[56] UC Berkeley (2021), Positive AI Economic Futures, World Economic Forum, https://www.weforum.org/reports/positive-ai-economic-futures.

[89] Ugale, G. and C. Hall (2024), “Generative AI for anti-corruption and integrity in government: Taking stock of promise, perils and practice”, OECD Artificial Intelligence Papers, No. 12, OECD Publishing, Paris, https://doi.org/10.1787/657a185a-en.

[145] UK Government Office for Science (2023), Future Risks of Frontier AI, https://assets.publishing.service.gov.uk/media/653bc393d10f3500139a6ac5/future-risks-of-frontier-ai-annex-a.pdf.

[98] UK National Audit Office (2023), Report on Accounts - Department for Work & Pensions, https://www.nao.org.uk/wp-content/uploads/2023/07/dwp-report-on-accounts-2022-23.pdf.

[318] UNESCO (2025), Artificial Intelligence and the Rule of Law, https://www.unesco.org/en/artificial-intelligence/rule-law.

[232] UNESCO (2019), Artificial Intelligence in Education: Challenges and Opportunities for Sustainable Development, https://repositorio.minedu.gob.pe/bitstream/handle/20.500.12799/6533/Artificial%20intelligence%20in%20education%20challenges%20and%20opportunities%20for%20sustainable%20development.pdf.

[316] United Kingdom (2023), Artificial Intelligence (AI) – Judicial Guidance, https://www.judiciary.uk/guidance-and-resources/artificial-intelligence-ai-judicial-guidance/.

[181] University of Turku (2023), TurkuNLP, https://turkunlp.org/ (accessed on 23 May 2024).

[274] UNU Institute for Environment and Human Security (2024), 5 Ways AI Can Strengthen Early Warning Systems, https://unu.edu/ehs/series/5-ways-ai-can-strengthen-early-warning-systems (accessed on 4 March 2025).

[264] US DHS (2024), DHS Launches First-of-its-Kind Initiative to Hire 50 Artificial Intelligence Experts in 2024, https://www.dhs.gov/archive/news/2024/02/06/dhs-launches-first-its-kind-initiative-hire-50-artificial-intelligence-experts-2024.

[263] US DHS (2024), United States Customs and Border Protection – AI Use Cases, https://www.dhs.gov/ai/use-case-inventory/cbp#deployed (accessed on 3 March 2025).

[147] US FOIA (2023), FOIA Reference Model White Paper, FOIA.gov, https://www.foia.gov/chief-foia-officers-council/foia-reference-model-white-paper-april-19-2023.

[258] US GAO (2024), Facial Recognition Technology: Federal Law Enforcement Agency Efforts Related to Civil Rights and Training, United States Government Accountability Office, https://www.gao.gov/products/gao-24-107372.

[44] Valderrama, M., M. Hermosilla and R. Garrido (2023), State of the Evidence: Algorithmic Transparency, https://www.opengovpartnership.org/documents/state-of-the-evidence-algorithmic-transparency/.

[208] van Baal, S. (2024), “Testing behaviour change with an artificial intelligence chatbot in a randomized controlled study.”, Journal Public Health, Vol. 45, pp. 506-522, https://doi.org/10.1057/s41271-024-00500-6.

[170] van Noordt, C. and G. Misuraca (2019), “New Wine in Old Bottles: Chatbots in Government”, in Lecture Notes in Computer Science, Electronic Participation, Springer International Publishing, Cham, https://doi.org/10.1007/978-3-030-27397-2_5.

[15] van Noordt, C. and L. Tangi (2023), “The dynamics of AI capability and its influence on public value creation of AI within public administration”, Government Information Quarterly, Vol. 40/4, p. 101860, https://doi.org/10.1016/j.giq.2023.101860.

[247] Vogl, T. et al. (2019), “Algorithmic Bureaucracy”, Proceedings of the 20th Annual International Conference on Digital Government Research, https://doi.org/10.1145/3325112.3325240.

[30] Walprecht, S. and C. Lewerenz (2024), Facilitating Regulatory Impact Assessments: The Benefits of Machine Learning in Legislation, https://www.destatis.de/EN/About-Us/Events/Machine-Learning/Slides/s2_walprecht.pdf.

[262] Wastupranata, L., S. Kong and L. Wang (2024), “Deep Learning for Abnormal Human Behavior Detection in Surveillance Videos—A Survey”, Electronics, Vol. 13/13, p. 2579, https://doi.org/10.3390/electronics13132579.

[154] Welker, Y. (2023), Generative AI holds great potential for those with disabilities - but it needs policy to shape it, https://www.weforum.org/agenda/2023/11/generative-ai-holds-potential-disabilities/ (accessed on 24 May 2024).

[157] Weyl, E., A. Tang and C. Plurality (2022), Augmented Deliberation, https://www.plurality.net/v/chapters/5-4/eng.

[11] Wisesa, A. (2023), “Inovasi artificial intelligence sebagai financial advisor kementerian keuangan”, Swatantra, Vol. XX, pp. 1-8, https://jurnal.umj.ac.id/index.php/SWATANTRA/article/viewFile/15611/8230.

[312] World Justice Project (2021), Grasping the Justice Gap: Opportunities and Challenges for People-Centered Justice Data, https://worldjusticeproject.org/news/grasping-justice-gap-opportunities-and-challenges-people-centered-justice-data.

[323] Xavier, O. et al. (2022), “Tax evasion identification using open data and artificial intelligence”, Revista de Administração Pública, Vol. 56/3, pp. 426-440, https://doi.org/10.1590/0034-761220210256x.

[275] Yao, M. (ed.) (2025), “Tailoring generative AI chatbots for multiethnic communities in disaster preparedness communication: extending the CASA paradigm”, Journal of Computer-Mediated Communication, Vol. 30/1, https://doi.org/10.1093/jcmc/zmae022.

[324] Zhang, G., H. Atasoy and M. Vasarhelyi (2022), “Continuous monitoring with machine learning and interactive data visualization: An application to a healthcare payroll process”, International Journal of Accounting Information Systems, Vol. 46, p. 100570, https://doi.org/10.1016/j.accinf.2022.100570.

[325] Zilka, M., H. Sargeant and A. Weller (2022), “Transparency, Governance and Regulation of Algorithmic Tools Deployed in the Criminal Justice System: a UK Case Study”, Proceedings of the 2022 AAAI/ACM Conference on AI, Ethics, and Society, https://doi.org/10.1145/3514094.3534200.

[105] Zinnbauer, D. (2025), Un-Plateauing Corruption Research? Perhaps less necessary, but more exciting than one might think, Elsevier BV, https://doi.org/10.2139/ssrn.5094531.

[265] Ziyadin, S. et al. (eds.) (2020), “Experience of Artificial Intelligence Implementation in Japan”, E3S Web of Conferences, Vol. 159, p. 04035, https://doi.org/10.1051/e3sconf/202015904035.

Notes

← 1.  https://www.ktpae.gr/erga/anavathmisi-kai-epektasi-upodomon-tpe-tomea-dikaiosinis.

← 2.  Information provided by Government of Greece officials.

← 3.  Information provided by officials of the Ministry of Presidency, Justice and Parliamentary Relations to the OECD.

← 4.  Protective measures refer to actions or legal mechanisms put in place to safeguard the rights, safety, or interests of individuals or groups, often in the context of vulnerable parties or during legal proceedings.

← 5.  Information provided by Government of Israel officials to the OECD.

← 6.  https://ministryofjustice.gr/?page_id=7483.