Artificial Intelligence in Italian Financial Markets

2. Approaches by Italian financial authorities to promote safe AI deployment

Abstract

This chapter examines the approaches adopted by Italian financial authorities to monitor AI deployment in the domestic market and to foster responsible innovation. The first section outlines existing supervisory measures and ongoing initiatives, including the use of Supervisory Technology (SupTech) tools. The second section highlights the innovation facilitator mechanisms that support an enabling environment for innovation and promote direct engagement between the authorities and the financial industry.

2.1. Introduction

Italian authorities have an active portfolio of tools and measures to monitor how AI technologies are being deployed in the domestic market. These include supervisory initiatives developed by the four supervisory authorities, in the form of data collection and research, as well as SupTech tools in production and development. These supervisory measures help the authorities to assess the scale of current AI deployment by financial institutions, map current trends and their impact, and project potential future developments in the field.

Currently, Italy has a well-developed ecosystem of innovation facilitators spanning all major segments of financial activity. These enable safe testing of AI applications in finance and foster constructive engagement with the industry. Innovation facilitators can play a valuable role in addressing regulatory barriers or gaps and sending a positive signal about the authorities’ commitment to responsible innovation.

2.2. Monitoring and supervising the deployment of AI applications in the Italian finance sector

Italian public authorities play an active role in promoting safe and responsible deployment of AI technologies within the domestic financial system. Support for innovation is evident in both the AI-related monitoring initiatives conducted by Italian financial supervisors and the operation of innovation facilitators, which enable direct engagement with the industry.

2.2.1. Monitoring and supervision initiatives by the Italian authorities

In Italy, the legislative and supervisory competences over the financial services are distributed between BdI, CONSOB, IVASS and COVIP. BdI exercises supervisory functions over banking and financial activities for banking, registered non-banking intermediaries and the payment services market (Banca d'Italia, 2024[1]). CONSOB is the supervisory authority for the Italian financial products market (CONSOB, 2024[2]). IVASS exercises supervisory authority over insurance and reinsurance undertakings, insurance groups, and other insurance intermediaries (IVASS, 2024[3]). COVIP supervises supplementary pension schemes (COVIP, 2025[4]). For certain issues (e.g. responsibility for financial stability), the competences are shared between the Ministry of Economy and Finance, BdI and CONSOB. Additionally, the MEF acts within cross-area competencies, including financial system regulation and policies, as well as customer protection, transparency and non-banking intermediaries (MEF, 2024[5]).

BdI, CONSOB and IVASS have undertaken several initiatives to collect information and data about the use of AI in the Italian financial sector, as listed in Table 2.1.

Table 2.1. AI-related initiatives of Banca d’Italia, CONSOB and IVASS

Banca d’Italia
AI-related data and information collection activities	FinTech Survey BdI conducts a Fintech Survey to assess the level of adoption of technological innovations across the financial services. The exercise occurs bi‑annually, with the first edition taking place in 2021, followed by the 2023 edition. The 2025 version features questions on GenAI and the implications of transposing the EU AI Act. In 2025, BdI launched a new survey on unregulated Italian fintech operators to supplement the previous survey. ICT self-assessment BdI releases templates for the assessment of the ICT risk with respect to the requirements set out in DORA. The self-assessment format contributes to a structured collection of information about the ICT risk across the supervised entities. The self-assessment of LSIs consists of a mandatory general section and a specific section required for only a specific sample of intermediaries. For the latter, one section is related to adoption of machine learning. Regional Bank Lending Survey (RBLS) Banca d’Italia collects significant data from the Regional Bank Lending Survey (RBLS) which examines, among others, relations between regional branches and their headquarters. This data is kept at aggregate level and is used solely for economic analysis, for example to examine the adoption of digital technologies in the banking sector (Banca d'Italia, 2022[6]). The survey has helped to identify the use of AI within FinTech projects, including its application in critical functions such as credit scoring, monitoring, AML and fraud detection. The 2024 edition of the survey marked the start of collection of data on GenAI, as well as the use of AI codes of conduct. The 2025 edition foresees collection of information on the role of BigTech companies in providing AI solutions, including the advancement stage of systems in their cycle. Inspections Although Banca d’Italia’s inspections are not yet specifically focussed on AI, the topic is assessed indirectly through related areas such as IT and credit risk management.
Research	Multidisciplinary team Banca d’Italia operates an internal multidisciplinary team on big data, machine learning and artificial intelligence, which includes economists, statisticians and computer scientists from different departments within the institution. Banca d’Italia has recently created a dedicated cross-departmental committee for the co‑ordination of some internal activities related to AI. AI in credit-scoring paper In 2022 Banca d’Italia published a paper to present a discussion of theoretical issues and of the regulatory and institutional context for the application of AI to credit scoring. The paper also describes the results of a survey of a selection of Italian financial intermediaries on their experience in adopting such models (Banca d'Italia, 2022[7]). Proofs of concept Exploring applications through proofs of concept. One of them, joint with some banks, showed that procedures based on AI solutions help reduce regulatory compliance costs for financial entities and verification costs for authorities Another one suggests that neural networks can support the detection of printing defects during banknotes production (Banca d'Italia, 2023[8]).
CONSOB
AI-related data and information collection activities	Data and information collection activities The supervision and monitoring activities on the use of AI applications are based on the use of unstructured information typically getting from the periodic supervisory documentation transmitted by the supervised entities, from any relevant inspection outcomes and, more in general, from the meetings held with intermediaries in the context of the supervisory activities. CONSOB collaborates in international groups (e.g. IOSCO, ESMA) and initiatives regarding the use of AI in the financial sector. In 2025, CONSOB contributed to the ESMA report on the AI adoption and trends in securities markets. CONSOB actively identifies, monitors, and assesses ICT-related risks (e.g. cyber risks) at both the individual firm level and systemic level. Its approach is based on supervisory rules (MiFID II and DORA). Changes in organisational structure and strategy In October 2024 CONSOB faced a re‑organisation intended, among other objectives, to enhance its capacity to steer initiatives in the field of artificial intelligence. As part of this process, two dedicated offices were established: the “Data Management and Artificial Intelligence”, dedicated to the development and management of business intelligence & advanced analytics platforms and AI solutions and models; the “Data Governance” Office, responsible for ensuring that the organisation’s data are managed in a proper, secure, consistent, and regulatory‑compliant manner. CONSOB’s 2025‑2027 Strategic Plan states that the Authority will strengthen supervisory efficiency by adopting AI‑based tools to improve transparency and oversight. The Plan highlights the development of advanced data‑driven supervision models, making data analysis central to Consob’s monitoring framework. It emphasises digital transformation as a transversal priority, with AI supporting more effective and timely supervisory actions. The Plan also supports the 2024 internal reorganisation reinforcing data and technology capabilities, including structures dedicated to AI and data management. Overall, AI and stronger data practices underpin a shift toward risk‑based, technologically enhanced market supervision throughout the 2025‑2027 horizon CONSOB is working on a “AI lifecycle management framework”, which provides a structured governance model that ensures AI initiatives align with regulatory requirements. It streamlines decision‑making by clarifying responsibilities across business units, the AI department, and external partners. The framework strengthens risk‑based prioritisation, enabling efficient allocation of resources and controlled deployment of AI solutions. It enhances transparency and traceability, supporting systematic approvals and more reliable operational outcomes. Overall, it helps the organisation adopt AI safely while maximising business value and strategic impact.
Research	A study on the development of the use of AI in the fields of wealth management, a survey conducted in 2021 with Assogestioni (CONSOB, 2022[9]). Survey response rate covered eight large asset management companies, accounting for 60% of assets under management in Italy. A study on “A machine learning approach to support decision in insider trading detection” – in which two complementary unsupervised machine learning methods to support market surveillance aimed at identifying potential insider trading activities are proposed (CONSOB - Scuola Normale Superiore di Pisa, 2022[10]). A study on “Dimensionally reduction techniques to support insider trading detection” (CONSOB, 2024[11]). An analysis on AI and market abuse (“AI and market abuse: do the laws of robotics apply to financial trading?”) (CONSOB, 2023[12]). “Insights on Artificial Intelligence and Financial Supervision”. This paper explores the challenges linked to the use of artificial intelligence systems in CONSOB’s market surveillance activities. (Deriu and Racioppi, 2025[13]) “Greenwashing alert system for EU green bonds. The CONSOB–University of Trento prototype” (Paterlini et al., 2025[14]). “The impact of the ESG factor on industrial performance. An analysis using machine learning techniques” (Palynska et al., 2024[15]). Sustainable Development Goals omission and environmental sentiment metric for greenwashing and ESG controversies alert in green bonds (Nicolodi et al., 2025[16]).
IVASS
AI-related data and information collection activities	IVASS is aware that the use of AI technologies is imbedded into insurance projects. It foresees the inclusion of AI in its general supervisory activities, comprising their overall performance and risks of discrimination. IVASS conducts its monitoring on AI with market surveys, one‑to‑one meetings with larger insurance undertakings and a continuous market outreach discussing with insurance and technology providers associations, academia, experts and other stakeholders. IVASS has not changed its organisational structure in any way to accommodate specificities of AI supervision as of yet.
Research	IVASS is currently conducting the following AI-related research: IVASS is conducting a project on network analysis, in collaboration with the University of Palermo. Statistical and ML techniques are used to analyse huge databases of insurance claims to identify groups of potential fraudsters (Tumminello et al., 2022[17]). IVASS undertook a qualitative‑quantitative analysis of the clarity of insurance contracts, in collaboration with the University of Torino. The projects used algorithms to assess the readability of insurance contracts, with the aim of ensuring clarity of the contracts, and the rights and obligations of consumers contained therein (IVASS, 2025[18]). IVASS contributed to a study conducted with EIOPA, for establishing more indicators of stability of insurance firms. A second stage of the project will feature GenAI technologies. IVASS conducted a survey on the use of Machine Learning algorithms by insurance companies in their relations with policyholders (IVASS, 2023[19]). IVASS published a working paper on AI governance in the insurance sector (Capone, 2021[20]).

Note: Non-exhaustive list.

2.2.2. Use of Supervisory Technology (SupTech) tools

SupTech tools can enhance the ability to track, audit, and challenge the performance and risks of AI systems used by financial institutions. These tools support data-driven oversight and improve the efficiency of supervisory activities. In a more complex world where NCAs responsibilities are increasing, SupTech initiatives are primarily aimed at strengthening supervisory capacity, analytical capabilities and risk prioritisation, rather than automating supervisory judgement or decision making.

Italian financial authorities are deploying SupTech tools in multiple areas of their supervisory activities. Banca d’Italia currently has six SupTech tools in production for the supervision of bank and non-bank financial institutions. CONSOB has also developed prototype SupTech tools spanning a range of areas. Table 2.2 lists the tools of the two authorities.

IVASS has also experimented with AI techniques on claims databases for fraud prevention to facilitate the consultation of insurance documents, to enrich indicators on the stability of insurance companies, and to classify complaints in order to make them easier to handle. Most projects have been developed in-house with the collaboration of third-parties (Banca d’Italia, EIOPA/ECB, academia). Analytical benefits of exploring big data bases include a possibility to define and test new sets of indicators, building internal skills and capabilities with a first-hand approach to practical projects. The main barriers to the development of SupTech tools are the availability of skilled staff, and the need to avoid conflicting priorities and to allow experimentation. Difficulties may also arise from the need to source external business intelligence, if required for developing specific internal SupTech applications. COVIP is also currently evaluating the possibility of developing SupTech tools.

Table 2.2. Use of SupTech tools by Banca d’Italia and CONSOB

Banca d’Italia
On-sight inspections	SupTech tools can enhance the efficiency of on-sight inspections by helping to automate routine activities i.e. to draft findings, perform consistency checks, analyse compliance documents and generate suggestions on the findings. The tool helps both on-site supervisors and consistency checkers. The architecture is composed of two application services responsible for, respectively providing the user with suggestions computed by AI about the finding that is being drafted/checked, as well as maintaining a structured and up-to-date data base of regulatory standards applicable to inspections.
Ongoing supervision	Ongoing supervision: Tools can enhance the ongoing supervisory processes by performing detailed analysis of the concrete areas of investigation, for example by reconstructing ownership chains in the financial sector. For corporate governance area, the tool supports the analysis of the effective functioning of board (behaviour of board members, in terms of type/frequency of interventions, depth of the discussion, number of topics discussed).
Fit and proper evaluations	The tool supports the Fit and Proper assessment by extracting information from various databases, such as: reporting on large exposures; reports on connected parties (both at individual and consolidated levels); Bankit Sanctions database; Italian Credit Register; Italian Company Register. Based on the extracted information, anomaly indicators are generated to guide the analyst in the decision-making process when assessing the suitability of financial institution representatives. Additionally, through an AI module based on the Factiva database, the tool enables the detection of reputational issues of the appointee by means of press articles.
Ownership structures	The tool reconstructs the ownership structure of supervised intermediaries (banks and financial companies) through the use of AI technologies, i.e. Automated Reasoning (AR) applied to their Knowledge Graph. This platform not only integrates and visualises data from multiple sources but also generates new knowledge beyond the scope of the original data, significantly enhancing the information available to users.
Consumer protection	In order to enhance the enforcement of consumer protection claims, BdI has been testing AI systems for analysing customer complaints (Esptech) and to support the role of the ombudsman (AbefTech). Both systems are developed internally. Esptech: The tool incorporating NLP and ML is used to manage over 10 000 complaints that are received annually. Models group the complaints in clusters based on semantic similarities in order to map trends and help identify the outliers. AbefTech: the tool supports the Banking and Financial Ombudsman’s Technical Secretariats for the management of complaints filed by customers through: a ML model to group the ABF’s decisions based on semantic similarities; a NLP function (Named Entity Recognition algorithms) to facilitate checks relating to the preliminary assessment of complaints (so-called admissibility); a function supported by combined techniques of Named Entity Recognition and regular expressions (which make use of the centralised registry of intermediaries present in the Institute) to facilitate the anonymisation of ABF decisions.
CONSOB
Market Surveillance/ Market abuse detection for insider trading:	Two complementary methods of unsupervised machine learning are used to identify potential insider trading activities: The first method “Pandora Box” uses clustering to identify discontinuities in an investor’s trading activity near a price‑sensitive event, such as a takeover bid. These discontinuities are compared to the investor’s past trading history and the present trading activity of their peers. The second unsupervised approach aims to identify small groups of investors who act coherently around price‑sensitive events. This may indicate potential insider rings, which are groups of synchronised traders displaying strong directional trading in a rewarding position in the period before the price‑sensitive event.
Fight against online harms	An AI solution aiming to fight against online harms, especially in relation to the illegal offering of crypto‑assets and investment services. The system is designed to automate the analysis and verification activities that currently characterise the complaint management process and a selected part of IOSCO online list: I-SCAN (International Securities Commodities Alert Network The technological core of the solution lies in the use of LLMs, including in their multimodal form, capable of processing textual and visual elements. The LLMs are able to understand complex texts, including legal and regulatory texts, and automatically extract structured data. The entire process includes a validation and back testing phase, in which the results produced by the system are compared with the analysis conducted by human operators. In the event of a misalignment, the system prompt is modified and the inference is repeated.
Risk-based analysis of issuers	A complex system is based on a combination of AI algorithms applied to the predictive analysis of the risks of listed issuers, to support supervision, and based on the list of risk parameters and both structured and unstructured data sources. The AI platform is based on data lake and uses ML algorithms that allow the identification of a cluster of risk issuers, identified through anomaly detection techniques. The AI platform also includes a trial based on a specific text classification algorithm, capable of labelling the text of the Key Audit Matters (KAM) indicated by the audit companies and transmitted to Consob, associating the text of each KAM with the respective international accounting standard of reference. The AI platform also includes a correlation of the analysed data with the applications of text-mining techniques for the screening of relevant information contained in unstructured data sources.
European green bonds supervision	A trial has been launched in collaboration with the University of Trento, incorporating the use of AI and NLP techniques. The core of the project, in this case, consists of building a computer prototype capable of automatically analysing the textual documentation associated with green bond issues, with the aim of identifying environmentally relevant phrases, labelling them according to their content, associating them with a sentiment rating and verify the presence or absence of references to sustainable development goals (SDGs). Subsequently, the AI system has been implemented through a technical engineering and an industrialisation process for cloud deployment. The added value of the model lies in its ability to generate “alerts” on documents and subjects with potentially anomalous characteristics – for example, a high mismatch between declared and actual SDGs, or a predominantly opportunistic sentiment in the absence of concrete evidence – which can guide the activities of the offices towards more in-depth investigations.
Sanctions anonymisation	The system is based on generative AI models and similarity algorithms, aimed at anonymizing the formal sanctions documentation by anonymisation of sanctions in a 4‑activities process, i.e. deleting from the text of the sanctioning measures the data and information potentially suitable for identifying the sanctioned party.

Source: Banca d’Italia and CONSOB.

2.3. Enabling environment for innovation in Italy

2.3.1. Financial regulatory sandbox

The financial regulatory sandbox was introduced by Ministerial Decree No. 100/2021, implementing Legislative Decree No. 34 of 2019. The sandbox allows supervised entities and FinTech operators to test innovative products and services, for a limited period of time, in constant dialogue with the three financial supervisory authorities (Banca d’Italia, CONSOB and IVASS), and eventually also benefitting from a simplified temporary regime. The testing activities are run by the three supervisors under the co‑ordination of the Fintech Committee, set up at the Ministry of Economy and Finance.

Within the first cohort of the sandbox supervised by the Bank of Italy, artificial intelligence was among the most widely adopted technologies in the projects tested. Out of the 13 projects admitted to the experimentation phase, 4 incorporated the use of artificial intelligence. IVASS is also currently overseeing a project which includes a limited use of AI, including for data analytics on driving behaviour in motor insurance and the generation of synthetic data for AI models. The use of AI has not been the focus of the initiatives analysed so far by CONSOB in the context of the sandbox, considering that fintech applications usually exploit several technologies, among which AI could be included.

After the first two cohorts, the Ministry of Economy and Finance and the three supervisors decided to review the secondary legislation regulating the sandbox (Ministry of Economy and Finance Decree No. 100 of 30 April 2021). The aim is to simplify access to the sandbox and reduce the administrative burdens on Fintech operators to encourage broader market participation. This will be achieved mainly by distinguishing the sandbox rules applicable depending on the type of testing, which varies according to market involvement and the potential risk to third parties. This ensures that operators can carry out their activities in a protected environment according to a risk-based and proportionate approach. The new version of the ministerial decree has been published and has been under consultation until 16 of May.

Article 57 of the AI Act requires that national competent authorities establish an AI regulatory sandbox either at the national level or jointly with other Member States (EU, 2024[21]). As anticipated by the MEF, discussions are ongoing to evaluate the extent to which the financial regulatory sandbox can be considered relevant even for the purposes of the regulatory sandboxes of the AI Act. This issue is subject to ongoing developments at the EU level, including the European Commission’s proposed “Digital Omnibus” package, which provides – among other things – for the establishment of an AI Sandbox at the EU level, without prejudice to national AI Sandboxes.

2.3.2. Other innovation facilitators

Milano Hub is an innovation centre created by Banca d’Italia in 2021 that aims to support the digital evolution of the financial market and attract talent and investment. Milano Hub offers consulting services, mentorship, and educational components to financial intermediaries, startups, and research centres, with the aim of accelerating the development of projects and promoting the quality and safety of specific innovations. In order to maintain a constant dialogue with market operators, Milano Hub works via its “Calls for Proposals” relating to different areas of innovation. The projects selected receive developmental support through: i) technical expertise in banking, finance and insurance and in specific areas, for example IT and regulation; ii) organisation of seminars, events, conferences with representatives from the projects, institutions and the academic world. BdI does not provide any form of financial assistance or any contribution to the acquisition of goods or services or support the projects in terms of promotional initiatives or marketing activities.

The First Call for Proposals was launched in 2021 under the theme “the contribution of artificial intelligence to improving the provision of banking, financial and payment services to businesses, households and the public administration, with a particular focus on financial inclusion, adequate consumer protection, and data security”. This call resulted in the submission of 40 projects by 62 entities, and the large majority of projects were related to digital lending & deposit (55%), payments (15%) and regtech (8%). The second Call for Proposals focussed on “the Application of Distributed Ledger Technology (DLT) to banking, financial, insurance and payment services”. This Call for Proposals resulted in 57 applications by 81 entities and 14 selected projects. The third Call for Proposals in 2024 focussed on “digital and instant payments”. Out of 26 projects submitted, 11 were selected. Among them, a few projects use AI algorithms to develop fast-payments for new use‑cases/applications dealing with fraud prevention techniques and cash-flow management systems (Banca d'Italia, 2024[22]).

FinTech Channel is the point of contact through which operators can dialogue easily and informally with Banca d’Italia. Operators can present projects in the field of financial services and of payments, based on innovative technology, or propose technological solutions designed for banks and financial intermediaries. In this way, start-ups, firms, banks and financial intermediaries proposing innovative solutions in the field of financial and payments services can contact the Bank. As of 4 April 2025, the Fintech Channel has had more than 260 interactions with fintech companies and financial intermediaries, with the percentage of projects focussed on AI solutions increasing from 20% to 33% in 2023, and to 38% in 2024.

References

[1] Banca d’Italia (2024), Bank of Italy - Supervision of the banking and financial system, https://www.bancaditalia.it/compiti/vigilanza/ (accessed on 16 December 2024).

[22] Banca d’Italia (2024), Le interconnessioni tra Intelligenza Artificiale, Cloud e Cyber nel settore finanziario.

[8] Banca d’Italia (2023), “Siamese neural networks for detecting banknote printing defects”, https://www.bancaditalia.it/pubblicazioni/mercati-infrastrutture-e-sistemi-di-pagamento/approfondimenti/2023-034/index.html?com.dotmarketing.htmlpage.language=1 (accessed on 25 April 2025).

[7] Banca d’Italia (2022), Bank of Italy - No. 721 - Artificial intelligence in credit scoring: an analysis of some experiences in the Italian financial system, https://www.bancaditalia.it/pubblicazioni/qef/2022-0721/index.html?com.dotmarketing.htmlpage.language=1 (accessed on 25 April 2024).

[6] Banca d’Italia (2022), No. 682 - The digital transformation in the Italian banking sector.

[20] Capone, D. (2021), “Governance of Artificial Intelligence in the insurance sector between ethical principles, board responsibility and business culture”, No. 16, IVASS, https://www.ivass.it/pubblicazioni-e-statistiche/pubblicazioni/quaderni/2021/iv16/index.html?com.dotmarketing.htmlpage.language=3&dotcache=refresh (accessed on 3 October 2025).

[2] CONSOB (2024), Activities - CONSOB and its activities - CONSOB, https://www.consob.it/web/consob-and-its-activities/activities (accessed on 17 December 2024).

[11] CONSOB (2024), “Dimensionally reduction techniques to support insider trading detection”, https://img.corrierecomunicazioni.it/wp-content/uploads/2024/03/01111218/94a3c409-0619-4d0b-9b67-1c9dcce90d3f.pdf (accessed on 21 May 2025).

[12] CONSOB (2023), “AI and market abuse: do the laws of robotics apply to financial trading?”, https://www.consob.it/documents/11973/201676/qg29en.pdf/62395177-4053-2549-5d44-b3e6d8be2408?t=1697185532182 (accessed on 21 May 2025).

[9] CONSOB (2022), Artificial intelligence in the asset and wealth management, https://www.consob.it/web/consob-and-its-activities/ft9en (accessed on 6 January 2025).

[10] CONSOB - Scuola Normale Superiore di Pisa (2022), A machine learning approach to support decision in insider trading detection, Quaderni FinTech, https://www.consob.it/documents/1912911/1933915/FinTech_11.pdf/eebb010d-e5e8-9f75-9e77-b2a1407e418f (accessed on 7 January 2025).

[4] COVIP (2025), About COVIP, https://www.covip.it/en/about-covip (accessed on 18 June 2025).

[13] Deriu, P. and S. Racioppi (2025), “Insights on Artificial Intelligence and Financial Supervision”, CONSOB, https://www.consob.it/web/consob-and-its-activities/ft15en (accessed on 14 October 2025).

[21] EU (2024), Regulation (EU) 2024/1689 (AI Act), https://eur-lex.europa.eu/eli/reg/2024/1689/oj (accessed on 28 November 2024).

[18] IVASS (2025), A qualitative-quantitative analysis of the clarity of insurance contracts: check of the progress, https://www.ivass.it/consumatori/azioni-tutela/indagini-tematiche/documenti/2025/follow-up-analisi-quali-quantitativa-contratti-febbraio-2025-en.pdf?language_id=3 (accessed on 4 March 2026).

[3] IVASS (2024), IVASS - Tasks, https://www.ivass.it/chi-siamo/in-sintesi/compiti/index.html (accessed on 17 December 2024).

[19] IVASS (2023), Survey on the use of Machine Learning algorithms by insurance companies in their relations with policyholders, https://www.ivass.it/pubblicazioni-e-statistiche/pubblicazioni/altre-pubblicazioni/2023/indagine-algoritmi/Esiti_indagine_Algogovernance_ENG.pdf?language_id=3 (accessed on 27 January 2025).

[5] MEF (2024), Department - MEF Department of Treasury, https://www.dt.mef.gov.it/en/dipartimento/ (accessed on 18 December 2024).

[16] Nicolodi, A. et al. (2025), “Sustainable Development Goals omission and environmental sentiment metric for greenwashing and ESG controversies alert in green bonds”, CONSOB Fintech Series No. 16, 2025, https://www.consob.it/documents/d/area-pubblica/fintech16 (accessed on 6 March 2026).

[15] Palynska, M. et al. (2024), “The impact of the ESG factor on industrial performance: An analysis using machine learning techniques”, CONSOB, http://www.consob.it (accessed on 14 October 2025).

[14] Paterlini, S. et al. (2025), “Greenwashing alert system for EU green bonds: The CONSOB-University of Trento prototype”, CONSOB, http://www.consob.it (accessed on 14 October 2025).

[17] Tumminello, M. et al. (2022), “Insurance Fraud Detection: A Statistically-Validated Network Approach”, https://www.ivass.it/pubblicazioni-e-statistiche/pubblicazioni/quaderni/2022/iv24/Quaderno_24.pdf (accessed on 4 March 2026).