AI adoption can improve public sector efficiency and service quality by supporting and accelerating administrative and support tasks.
Building internal capability for AI is important in public administrations to ensure compliance, accountability, and effectively leverage AI tools to achieve organisational goals.
Institutions need a proactive and robust governance of AI to prevent significant risks, also related to the use of generative AI tools by staff.
AI adoption will change work processes and skills needed within public administration. Investing in training and upskilling can help people and institutions adapt to these changes.
More and more institutions develop training on AI for their staff, targeting all employees to develop foundational AI skills, leaders to gain strategic knowledge on AI, and digital and data professionals to improve technical AI-related skills. Examples include the “AI Masterclass for Senior Leaders” in Ireland and the “ABC of AI” e‑learning module in Estonia.
Building an AI‑ready public workforce
Public administrations are major employers in OECD countries and many of them are facing staff shortages, high workloads, and fiscal pressures. Among their core functions are administrative and support tasks, for example, processing documents, managing claims, or providing information to individuals. AI systems can support and accelerate these procedures, improving service quality and freeing up staff capacity for more complex tasks. At the same time, a lack of skills and internal capability is among the most widely cited barriers to the adoption of AI in the public sector. Preparing people, in this context, is as important as deploying technology.
This policy brief explores the implications of AI adoption for the workforce in public administrations, and outlines strategies public sector leaders can pursue to strengthen AI capabilities of their staff and within their institution. It builds on the OECD report Harnessing AI in Social Security: Use Cases, Governance, and Workforce Readiness (OECD, 2025[1]), and draws on examples of workforce initiatives from various OECD countries, highlighting approaches towards a trustworthy and effective use of AI.
AI systems have significant potential to boost efficiency and service quality in public administration. The technology may help reduce administrative burdens, streamline service provision, and free up capacity in public administration. Particularly rule‑based administrative procedures – across different areas of government – may be organised into different components that can be supported by AI solutions. For example, Kela, Finland’s national social security institution uses an AI platform to automate the classification and processing of documents attached to benefit applications, saving an estimated 38 years of full-time equivalent (FTE) work for case workers per year.
Governments carefully scrutinise the internal adoption of new technology, assessing costs, benefits, risks, regulatory considerations, and workforce impact. They proceed with caution to identify use cases and develop pilots where AI technology adds value, and so far, AI adoption in public administration has been limited. Human oversight, transparency, accountability and trust continue to be central principles for government procedures and services. Concerns that AI will worsen the quality of jobs or replace employees in the public sector are currently speculative.
The use of generative AI (genAI) tools in public administration requires proactive and robust governance. Surveys indicate that a significant proportion of public servants already use open-access genAI tools such as ChatGPT to support their work. If used in ways that conflict with organisational policies and regulations, however, these tools can pose risks, for example, to data protection. Advanced public sector organisations have responded with staff training, clear rules and governance around AI use, and internal genAI tools. For instance, the municipality of Gladsaxe in Denmark operates a self-hosted large language model (LLM) based on GPT technology.
Capability gaps are a barrier to AI adoption, and institutional skill needs are changing over time. Internal skills gaps related to AI are a widespread challenge in government institutions. Going forward, digital and data professionals with advanced technical skills will play a key role in developing tailored and integrated AI solutions. At the same time, most employees will need “AI literacy” to use and interact with AI tools effectively and responsibly. Beyond technical skills, complementary skills such as problem-solving or critical thinking, and transversal skills such as social, management, or teamwork skills are becoming increasingly important.
Strategic AI use in government will transform work processes. Realising the potential of AI systems in public administration implies changes to how work is organised and how institutions deliver value. Rather than automating existing processes, AI can be part of a redesign towards integrated, citizen-centred public services. For public sector employees, this transformation means that they need to be open to change, continuously learn, and explore innovative applications of AI to achieve institutional goals.
Governments have three main levers to strengthen their AI capabilities: outsourcing, hiring, and training. Procuring certain tasks to third-party providers is a standard practice across governments, however, it must be balanced with the need for accountability. Building in-house AI capability rather than outsourcing tasks helps to develop AI systems that are in line with institutional needs, prevents information asymmetries in procurement, and avoids dependencies on technology providers. This approach also requires aligning workforce development with the broader institutional strategy on AI, data and technology adoption.
Strategic GovTech collaborations can help strengthen AI capability. Governments often partner with start-ups, innovative SMEs, research centres, or universities. These collaborations can take various forms, such as centres of public sector innovation within government institutions or external non-profit association with a facilitating role. Common activities include open competitions, hackathons, accelerator programmes, pilot experiments, research and development grants, and peer learning networks. Unlike standard outsourcing, GovTech collaborations offer an ecosystem to test new technology solutions.
Targeted talent programmes help attract and retain digital and data professionals. Recruitment of scarce ICT professionals in public administration is a challenge, but some governments offer competitive programmes for high-impact technology projects, such as https://beta.gouv.fr/ in the French public administration and the Presidential Innovation Fellows in the United States. These programmes aim to attract high-skilled applicants from both the public and private sectors for time‑bound digital transformation projects. Other initiatives focus on improving hiring procedures, remuneration, and career pathways for digital and data professionals. An example is the “Digital and Data Profession” in the Australian Public Service, which promotes recruitment, career development and training opportunities for staff in these roles.
Public institutions can foster in-house innovation and continuous learning on AI through various levers. They should deliberately create an environment that encourages and stimulates curiosity, critical thinking, and experimentation to support AI adoption and digital transformation. The following approaches are part of the “toolbox” to foster an organisational culture of innovation and continuous learning:
Spaces for experimentation in courses or labs
Regular and open events on AI for all staff
Communities of practice
Multidisciplinary innovation teams
Innovation competitions
Regular performance reviews
The City of Helsinki hosts a competition called Kokeilukiihdyttämö, or Experimentation Accelerator, to foster and reward innovation on AI and digital technology among its 38 600 employees. Part of the city’s digital transformation programme, the accelerator has to date supported 65 internal experiments with AI technology and new work methods. Each selected proposal receives EUR 10 000 enabling teams to work with third-party providers and develop the suggested technology solution. Experiments aim not only at implementation but also to generate valuable “lessons learnt”, which are documented on a dedicated webpage for future projects.
Source: City of Helsinki (2025[2]), Kokeilugalleria [Experimental gallery], https://kokeilukiihdyttamo.hel.fi/results.
AI training is becoming increasingly common in public administration. The relatively high level of job security in the public sector presents both an opportunity and an obligation for institutions to invest in and incentivise their workforce’s skills development. Public administrations need to ensure that staff can use AI tools effectively and responsibly and avoid that some employee groups are left behind in the technology transition. In the European Union, organisations that provide or deploy AI systems are legally required to ensure their staff has a “sufficient level of AI literacy”, according to Article 4 of the AI Act.
Most institutions in the public sector lack a structured approach to identifying AI-related skills needs. Some use DigComp, the European reference framework for digital skills, which also covers AI-related skills.
Based on the AI Skills for Business Competency Framework by the Alan Turing Institute (2024[3]) and the OECD Framework for Digital Talent and Skills in the Public Sector (OECD, 2021[4]), three key workforce groups with different skills needs for AI adoption can be identified:
General employees need to be able to interact with and use AI tools effectively and responsibly. Next to basic digital skills, they require general knowledge and awareness on emerging AI technology, its potential, risks and ethical use, awareness of the need for data protection, and critical thinking and independent judgement.
Leaders require a strategic understanding of AI technology, including its potential and risks for their organisations’ objectives. This entails knowledge on emerging AI technology, the capacity to consider risks, ethical, and regulatory aspects related to AI, strategic thinking for AI implementation and capability building on AI governance, data infrastructure, and workforce readiness, and strong collaboration, communication, stakeholder management and change management skills.
Digital and data professionals need the skills to support the development, implementation and maintenance of AI systems. Besides technical skills in their field of specialisation such as data science, they need detailed knowledge on ethical and regulatory considerations around AI, technical skills to mitigate risks and manage the complexity of AI systems, as well as collaboration and communication skills to work in interdisciplinary teams and support leadership and the organisation.
Many government institutions provide foundational training on AI for employees. This training is often delivered in the format of short online courses, such as the “ABC of AI” e‑learning by the Digital State Academy of Estonia. Similarly, the French Caisse Nationale d’Assurance Maladie (CNAM) offers a self-paced one‑hour e‑learning module to familiarise staff with the main concepts of genAI, understand its opportunities and risks, and adopt responsible, effective practices. The Canada School of Public Service’s Digital Academy offers an introductory AI course as a virtual classroom-based training, delivered iteratively to cohorts of civil servants.
Training for leaders often focusses on the strategic potential of AI for their organisation. For example, Civil Service Learning in the UK’s Government Digital Service offers an online training course on the business value of AI. The Institute of Public Administration (IPA) in Ireland offers an interactive in-person training called “AI Masterclass for Senior Leaders” to help senior managers in the public service enhance their understanding of AI and its strategic potential for policymaking and service delivery.
Training for digital and data experts typically covers AI regulation and ethics, specific use cases and technical skills. The City of Helsinki, for instance, provides periodic training on data science and data engineering based on staff needs. GovTech Germany, a government-funded non-profit organisation, offers a four‑hour in-person training on genAI for human-centred public services based on concrete use cases. Another example is the interactive two‑day training for digital professionals on ensuring compliance with the EU AI Act offered by the European Institute of Public Administration.
There is a trade‑off between the number of participants and the cost of training (see Figure 1). Under resource constraints, shorter online training can easily be scaled to many participants, while more intensive and costly courses are typically offered to a selected target group. Training formats range from short online courses, structured training programmes, to modular programmes that lead to a certification. In Austria, for example, the Federal Administrative Academy offers a certificate for participants who have completed eight days of training on AI, digital, or data-related topics within three years. Going forward, AI technology could support the development and implementation of customised training itself, even if to date such applications are still rare.
Source: OECD (2025[1]), Harnessing Artificial Intelligence in Social Security: Use Cases, Governance and Workforce Readiness, https://doi.org/10.1787/b52405c1-en.
Training is more effective when facilitated by a trainer and tailored to the work context. Evidence suggests that trainer-led courses, whether in-person or online, are more effective than self-paced ones. Training should be tailored to the work context and include practical applications rather than only provide information (Schuster and GOV.UK, 2024[5]). Longer-term effectiveness of training interventions may be supported by a conducive environment for learning and innovation, for example, through performance reviews, innovation competitions, or communities of practice. Measuring the impact of training is good practice to understand a training’s return on investment and improve its content and design over time.
Build in-house AI capability. Combine targeted hiring, strategic GovTech partnerships, and training initiatives to build in-house capability for AI. This approach ensures accountability and aligns AI implementation with institutional objectives.
Invest in training for all staff groups. Provide employees with foundational knowledge and clear guidelines on AI use. Equip digital and data professionals with skills to develop, implement, and maintain AI systems in line with ethical and regulatory standards. Ensure leaders have a strategic understanding of AI’s potential and risk for their organisation.
Foster innovation and continuous learning. Create environments that encourage experimentation and innovation, for example, through competitions, communities of practice, or dedicated innovation teams.
Align workforce development with the strategy on AI and data. Support changes in work organisation through forward-looking workforce development and AI capability building. Align these efforts with the institution’s technology strategy to realise the potential of AI.
OECD (2025), Harnessing Artificial Intelligence in Social Security: Use Cases, Governance and Workforce Readiness, OECD Digital Government Studies, OECD Publishing, Paris, https://doi.org/10.1787/b52405c1-en.
References
[2] City of Helsinki (2025), Kokeilugalleria [Experimental gallery], https://kokeilukiihdyttamo.hel.fi/results.
[6] OECD (2025), Governing with Artificial Intelligence: The State of Play and Way Forward in Core Government Functions, OECD Publishing, Paris, https://doi.org/10.1787/795de142-en.
[1] OECD (2025), Harnessing Artificial Intelligence in Social Security: Use Cases, Governance and Workforce Readiness, OECD Digital Government Studies, OECD Publishing, Paris, https://doi.org/10.1787/b52405c1-en.
[4] OECD (2021), “The OECD Framework for digital talent and skills in the public sector”, OECD Working Papers on Public Governance, No. 45, OECD Publishing, Paris, https://doi.org/10.1787/4e7c3f58-en.
[5] Schuster, C. and GOV.UK (2024), What do we know about the design of effective management training in government? A rapid evidence review to inform training design and evaluation, https://www.gov.uk/government/publications/effective-management-training-in-government/what-do-we-know-about-the-design-of-effective-management-training-in-government-a-rapid-evidence-review-to-inform-training-design-and-evaluation (accessed on 8 October 2025).
[3] The Alan Turing Institute (2024), AI Skills for Business Competency Framework, https://zenodo.org/records/11092677/preview/AISkillsForBusinessCompetencyFramework_V2.pdf?include_deleted=0 (accessed on 2 June 2025).
Contact
Magdalena Burtscher (✉ magdalena.burtscher@oecd.org)
This policy brief was developed in the context of a project that was funded by the European Union via the Technical Support Instrument, and implemented by the OECD, in co‑operation with the European Commission.