Record STMT-W249585Voluntary

Artificial Intelligence Playbook for the UK Government

Guidelines Institutional policyNorthern Europe · GBadopted

The AI Playbook for the UK Government provides updated and comprehensive guidance for public sector organisations on harnessing the power of artificial intelligence safely, effectively, and responsibly. Developed collaboratively across government, academia, and industry, it outlines ten core principles, practical implementation steps, ethical considerations, security requirements, and governance structures. The playbook covers various AI technologies—including machine learning, natural language processing, and generative AI—to help civil servants improve public services while managing risks like bias, data leakage, and hallucinations.

Summary and content fields are drawn from the document text and are pending editorial review.

Who issued it

Department for Science, Innovation and TechnologyGovernment

Lead issuer

Legal forceVoluntary

Under whose authorityInstitutional policy

Judged fromfull text

Who it governs

National governmentRegulatorsPublic sector bodiesAI developers and providersAI deployers and usersResearchers and academics

This playbook is for UK civil servants, public sector leaders, policymakers, technical teams, and commercial professionals designing, procuring, or deploying AI solutions.

Sectors

Public administration · uk public sector

What the document contains

What's covered

The playbook covers ten foundational principles for AI use in government, primers on various AI fields and generative AI, practical steps for building AI solutions, user research methodologies, procurement routes and requirements, ethical and legal considerations including data protection and equality, security risks and mitigations, and governance structures like AI review boards and systems inventories.

What's missing

While the playbook provides comprehensive guidance for public sector adoption, it does not replace specific organisational policies or provide legally binding binding rules for private sector entities outside government supply chains.

Key takeaways

Public sector organizations must follow ten core principles to ensure AI is used lawfully, ethically, securely, and with meaningful human oversight.
The playbook expands beyond generative AI to encompass all current fields of artificial intelligence, providing primers and learning resources for civil servants.
Multidisciplinary teams incorporating technical, legal, commercial, and user research expertise are essential for successful and responsible AI deployment.
Rigorous risk management, data protection impact assessments, and quality assurance techniques must be integrated throughout the entire AI lifecycle.
Government bodies must implement robust cybersecurity measures and be vigilant against unique AI threats such as prompt injections, data poisoning, and model hallucinations.

Principles it states

Principle 1: You know what AI is and what its limitations arePrinciple 2: You use AI lawfully, ethically and responsiblyPrinciple 3: You know how to use AI securelyPrinciple 4: You have meaningful human control at the right stagesPrinciple 5: You understand how to manage the full AI life cyclePrinciple 6: You use the right tool for the jobPrinciple 7: You are open and collaborativePrinciple 8: You work with commercial colleagues from the startPrinciple 9: You have the skills and expertise needed to implement and use AI solutionsPrinciple 10: You use these principles alongside your organisation’s policies and have the right assurance in place

Outline

· Acknowledgements
· Update log
· Foreword
· Preface
· Principles
· Understanding AI
· Building AI solutions
· Working collaboratively
· Using AI safely and responsibly
· Security
· Governance
· Appendix: example AI use cases in the public sector

Issued1 February 2025

Length36,313 words

FormatNot recorded

Scopenational

How it scores

Pending review

Each score is backed by a quote verified against the document text. Editorial review has not yet confirmed them.

Accountability

4/5

Clear responsibility for AI outcomes

Ultimately, responsibility for any output or decision made or supported by an AI system always rests with the public organisation.

Accountability and responsibility

screening score · 2026-10-23

Fairness

4/5

Non-discrimination and equitable treatment

Implement bias mitigation and fairness evaluation across the entire AI project life cycle. Comply with human rights law, the Equality Act 2010

Fairness, bias and discrimination

screening score · 2026-10-23

Human Dignity

2/5

Respect for the inherent worth of persons

You must ensure that AI systems generate a positive impact on stakeholders and civil society at large while minimising potential harms as much as possible.

Principle 2: You use AI lawfully, ethically and responsibly

screening score · 2026-10-23

Human Oversight

4/5

Humans remain in control of consequential decisions

This includes ensuring that humans validate any high-risk decisions inﬂuenced by AI and that you have strategies for meaningful intervention.

Principle 4: You have meaningful human control at the right stages

screening score · 2026-10-23

Privacy

4/5

Protection of personal data and private life

AI systems can process personal data, so you need to consider how you protect this personal data, be compliant with data protection legislation

Principle 2: You use AI lawfully, ethically and responsibly

screening score · 2026-10-23

Transparency

4/5

Openness about how AI systems work and are used

If you’re a central government department or an arm’s length body within scope, you’re required to use the Algorithmic Transparency Recording Standard (ATRS) .

Principle 7: You are open and collaborative

screening score · 2026-10-23

Enforcement

4/5

Consequences and compliance mechanisms with real force

All central government departments and their arm’s length bodies are expected to follow this on a ‘comply or explain’ basis.

Existing guidance

screening score · 2026-10-23

Impact Assessment

4/5

Required assessment of an AI system's risks/impacts before or during use

Before implementing AI solutions, you need to undertake a data protection impact assessment (DPIA) . This involves an assessment of data protection and privacy risks

Lawfulness and purpose limitation

screening score · 2026-10-23

Topics

Physical AI Governancecore

artificial_intelligence public_sector ai_governance ethics cyber_security procurement