AI Roadmap
United Kingdom · Guideline · AI Council
Our summary
An independent report from the AI Council that outlines strategic recommendations for the UK government's approach to artificial intelligence. It identifies four foundational pillars to position the United Kingdom as a leader in safe and responsible AI development, calling for coordinated action across government departments.
Written by the Index in our own words. It is not legal advice; the official text is the authority.
Inside this policy
Every section of this policy in order. Our notes summarise each one in plain language and do not quote the official text, which is the authority.
Drafted by the Index An automated summary of the official text, not yet reviewed by a person.Reviewed Checked by a person against the official text.
13 sections
Jump to a chapter
UK AI council AI roadm ap
The AI Council is an independent expert body offering direction to the UK Government on the artificial intelligence sector. This report identifies four foundational areas for building the UK's position in AI and calls for government action to ensure safe and responsible AI development.
UK AI council | AI roadmap
The roadmap sets out 16 recommendations under four pillars - backing investigation and innovation, expanding population-wide technical understanding and capabilities, establishing trusted digital and physical infrastructure with flexible rules and effective oversight, and adopting sector-wide approaches to deploy AI - to position the UK as a leading place to develop and work with AI by ensuring public confidence in the science, technologies, governance and regulation. Recommendations include scaling public investment in AI, committing to a 10-year skills programme, building infrastructure for data access, supporting AI capability across companies and sectors, promoting diversity, establishing an online academy for AI literacy, developing governance standards, ensuring public scrutiny of automated decision-making, supporting AI startups, and enabling AI applications in climate change, defence and security, and healthcare. Government action should strengthen the R&D Roadmap to support cutting-edge research and cross-disciplinary work, attract and retain top AI talent, and ensure funding supports both basic and applied research; and the Alan Turing Institute should be enhanced and reshaped as crucial research, development and innovation infrastructure with full national reach across the four nations, establishing regional and local centres-of-excellence or hubs, providing skills training at research level and through apprenticeships and reskilling programmes, and requiring assured long-term public sector funding.
Research, development and innovation 13
The Roadmap supports goal-directed moonshot programmes in an ARPA-esque style to advance AI development through challenges to develop safe, ethical, explainable and reproducible AI methods and to explore human-AI partnerships, with AI integrated into moonshots across sectors such as materials science and electricity distribution for achieving net zero emissions. Moonshot projects should inspire public and business engagement with emerging AI science, enable exploration of human-AI partnerships, and develop responsible and ethical practices, while challenge-based funding programmes should select mature sectors ready for AI adoption and use mission-based approaches to integrate AI across disciplines and research horizons. Forward-looking frameworks could consist of labs building engineered prototypes and platforms to generate vision-based projects that are flexible models of experimentation.
Research, development and innovation 15
The UK must continue investing in basic AI research to address fundamental challenges including common data science languages and dataset interoperability, broaden AI systems beyond narrow disciplinary bases to diverse domains and unstructured data, and maintain global research leadership by keeping pace with AI's disruptive effects across all disciplines including social sciences and humanities. All children should leave school with basic understanding of how AI works, including what questions to ask and ethical implications beyond coding or mathematics. The UK government should commit to scaling up graduate-level AI education significantly across apprenticeships, Masters, Centres for Doctoral Training and PhDs at all levels for at least a decade, extend AI courses beyond traditional data science disciplines, and ensure no graduate leaves university without grounding in AI.
Skills and diversity 19
Teacher support should extend the NCCE's work with curriculum resources on AI at all Key Stages and incentivise teachers to develop AI knowledge through learning that combines knowledge with real-world problem-solving, curated datasets, and collaboration between universities, colleges and businesses with mentoring and data sharing available across the country to young people of all backgrounds. The government should extend business tax relief to investment in specified retraining and upskilling including AI, create an Online Academy for AI providing entry-level material, teacher resources, product manager support, and a data skills observatory, enable public engagement and dialogue on responsible AI innovation, and connect people to research through the Turing. The UK should become the world's leading place for safe, secure and high-quality data access and use for developing AI applications and public engagement with automated decision-making systems, with the government creating an AI-friendly Information Management Framework for clean real-time data, funding data infrastructure as a national asset, and private sector regulators helping businesses deploy privacy-protecting technologies and establish frameworks governing data sharing through trusts, cooperatives and contracts.
Data, infrastructure and public trust 23
Strong data governance frameworks and adaptive regulation are essential to build public trust in AI; independent oversight is needed to address public scepticism about safety, security, fairness and ethics, and to reconcile the mismatch between market logic extracting value from collective data and laws addressing individual rights. Methods should complement individual rights-based approaches by ensuring public scrutiny of automated decision-making, transparency about algorithm purposes and training data, and meaningful public input with ability to enforce sanctions; the Office for National Statistics should lead developing governance frameworks for data generated by individuals in daily activities, and an independent entity should provide recommendations on evolving governance mechanisms including impact and risk assessments, ethical processes and institutional mechanisms. The UK should foster responsible practices among developers and organisations while maintaining adequately-funded regulatory institutions with expertise in data protection and algorithmic oversight, merging sector-specific knowledge with cross-sector perspectives to bring together different regulators.
Data, infrastructure and public trust 25
Proposes developing digital twins with interoperable data standards to enable safer and cheaper testing of AI applications before deployment in the physical world, and establishing a nationwide network of living laboratories as controlled physical environments where AI systems can be safely tested and demonstrated with real-world data. Recommends the UK leverage its research strength and diplomatic influence to shape international AI discussions through partnerships with like-minded countries and leadership in forums such as the Global Partnership for AI, and to invest strategically in AI research and innovation to reduce dependence on critical technologies from non-aligned sources.
Data, infrastructure and public trust 27
Recommends encouraging business adoption of AI solutions by building confidence through impartial advice and robust products, particularly for SMEs, and supporting the AI startup and scale-up ecosystem through initiatives like Tech Nation's Applied AI Program and potentially increased venture capital funding or public investment in companies serving government priorities. Suggests assisting startups and scale-ups to recruit talent by extending study visas and positioning the UK as a leading AI startup location, introducing AI-specific Knowledge Transfer Partnerships to commercialise research, and promoting skilled product managers to the UK, while enabling public sector adoption of AI through data access mechanisms across government organisations and outcome-based projects.
National, cross-sector adoption 31
Government should enable public sector AI adoption through increased data science training, interdisciplinary projects linking sectors and departments, strategic leadership, improved procurement approaches, independent ethics committees to protect vulnerable populations, and cross-sector learning and integration of public and private sector data. The healthcare sector should lead on AI adoption to manage complex information flows between health, mental health and social care systems with NHSX taking a lead role, including developing a health data strategy with strong governance, creating partnership models between SMEs and the NHS, building skills for healthcare professionals and citizens, and using AI to advance biomedical research and clinical trials.
National, cross-sector adoption 33
AI is essential for reducing greenhouse gases and environmental impacts and supporting defence and security decision-making, requiring the UK to develop a plan for responsible AI in energy and climate, encourage open international data sharing on environmental information, and use its cleantech capabilities. AI techniques must be scaled and operate in partnership with human personnel to ensure decisions are understandable and justifiable, and are needed to monitor complex networked systems, detect and respond to cyber-attacks, and provide secure intelligence. The UK should strengthen ties with allied nations to develop advanced technologies together and increase investment in competitive hardware technologies.
National, cross-sector adoption 35
Suggests joint development of technology innovations across different sectors, without restricting to particular applications, as a potential approach to strengthen Britain's technological competitiveness and attract leading scientists and technologists.
Blank page
Describes the AI Council as an independent expert committee providing advice to the UK Government on artificial intelligence, with the Office for Artificial Intelligence serving as its secretariat.
Beis dcms
Controls
What this policy requires, forbids, grants or penalises, one control at a time and in our own words. Each control links to the official text, which is the authority. This is not legal advice.
Drafted by the Index Drafted automatically from the official text, not yet reviewed by a person.Reviewed Checked by a person against the official text.
Themes covered 13 of 35
Choose a theme to see only its controls. About the themes
Risk and scope 1 of 4
- Risk classification, not covered
- Prohibited practices, not covered
- Impact assessment, not covered
Data and privacy 2 of 4
- Cross-border data transfer, not covered
- Copyright and training data, not covered
Transparency and records 1 of 4
- Labelling of synthetic content, not covered
- Technical documentation, not covered
- Record keeping and logging, not covered
Technical safeguards 3 of 4
- Bias and non-discrimination, not covered
Assurance and lifecycle 0 of 4
- Conformity assessment and certification, not covered
- Registration and notification, not covered
- Incident and breach reporting, not covered
- Post-market monitoring and corrective action, not covered
Rights and protected groups 0 of 4
- Automated decision-making rights, not covered
- Children and vulnerable groups, not covered
- Biometric identification and surveillance, not covered
- Whistleblowing, complaints and redress, not covered
Roles and organisation 5 of 6
- Supply chain and provider-deployer duties, not covered
Enforcement and sectors 1 of 5
- Enforcement, penalties and regulators, not covered
- Sector rules: finance and insurance, not covered
- Sector rules: employment, not covered
- Sector rules: education, not covered
The full list of controls could not be loaded, so only the first 29 are shown.
UK AI council | AI roadmap 3 controls
- ObligationApplies to: Public authorityDrafted by the Index
The government should develop a National AI Strategy prioritising research, skills, data and infrastructure, and cross-sector adoption to position the UK for success in AI.
- ObligationApplies to: Public authorityDrafted by the Index
The government should scale up investment in AI over a decade and beyond to build the UK's competitive position and attract talent.
- ObligationApplies to: Public authorityDrafted by the Index
The government should maintain public confidence in AI science, technology and governance to realise economic and societal benefits.
Research, development and innovation 13 4 controls
- ObligationApplies to: Public authorityDrafted by the Index
The government should provide sustained public funding to a well-funded national research institute so it can plan for the medium term and develop strategic leadership in AI.
- ObligationApplies to: Public authorityDrafted by the Index
The UK should establish itself as a centre of expertise in explainable and trustworthy AI development through its research institutions and collaborations.
- ObligationApplies to: Public authorityDrafted by the Index
The government should fund challenge-led innovation programmes in the style of ARPA initiatives to advance next-generation AI developments such as learning from small data sets and novel neural architectures.
- ObligationApplies to: Public authorityDrafted by the Index
The government should support development of methods and approaches for safe, ethical, explainable and reproducible AI to accelerate its deployment across many sectors.
Research, development and innovation 15 2 controls
- ObligationApplies to: OthersDrafted by the Index
Researchers should continue and expand investment in basic research into AI fundamentals across multiple disciplines to address challenges like data interoperability.
- ObligationApplies to: Public authorityDrafted by the Index
The UK should ensure the AI workforce reflects societal diversity and that everyone can live confidently with AI.
Skills and diversity 19 4 controls
- ObligationApplies to: Public authorityDrafted by the Index
The government should develop and expand an international fellowship programme to recruit high-calibre researchers and build momentum in UK AI talent.
- ObligationApplies to: OthersDrafted by the Index
Providers of education should create and distribute comprehensive curriculum resources covering AI at all Key Stages and support teachers in learning and developing their knowledge of AI.
- ObligationApplies to: AnyoneDrafted by the Index
Educational and commercial organisations should collaborate to ensure graduates and apprentices can enter emerging AI professions, with learning delivered across all regions and to young people from all backgrounds.
- ObligationApplies to: OthersDrafted by the Index
Educational institutions should provide curriculum resources that include curated datasets to help teachers create case studies and exercises that build learner confidence and data literacy.
Data, infrastructure and public trust 23 3 controls
- ObligationApplies to: AnyoneDrafted by the Index
A coalition of institutions and experts should develop standard data sharing agreements that comply with data protection and intellectual property law and can be adapted across the economy.
- ObligationApplies to: Public authorityDrafted by the Index
The government should implement adaptive and informed regulation, standards and public engagement measures as part of a governance system that enables AI innovation while building public trust.
- ObligationApplies to: Public authorityDrafted by the Index
The government should explore mechanisms such as a public interest data bill to increase accountability and public engagement in how data is used for AI and algorithmic applications.
Data, infrastructure and public trust 25 2 controls
- ObligationApplies to: Public authorityDrafted by the Index
A cross-sectoral UK Digital Twin programme should be established to create realistic representations of assets and systems with interoperable data standards, enabling safe exploration of AI applications with synthetic data.
- ObligationApplies to: Public authorityDrafted by the Index
The UK should create a nationwide network of living laboratories as physical environments where AI systems can be tested in constrained real-world settings to explore safe innovation without causing inadvertent harm.
Data, infrastructure and public trust 27 3 controls
- ObligationApplies to: Public authorityDrafted by the Index
The government should encourage initiatives to support high-growth AI startups and scale-ups through programs like Tech Nation's Applied AI Program.
- ObligationApplies to: Public authorityDrafted by the Index
The government should provide physical infrastructure and support to enable AI startups and scale-ups to scale, develop and serve local and multinational companies.
- ObligationApplies to: Public authorityDrafted by the Index
Grants and impartial advice should be provided to SMEs to help them adopt digital and AI solutions for business growth.
National, cross-sector adoption 31 5 controls
- ObligationApplies to: Public authorityDrafted by the Index
Government departments should train data analysts and AI practitioners in data science techniques and collaborate with research organisations to establish professional standards.
- ObligationApplies to: Public authorityDrafted by the Index
The government should fund interdisciplinary pathfinder projects involving collaboration between industries, government departments and academic disciplines to demonstrate and build capability in AI deployment.
- ObligationApplies to: Public authorityDrafted by the Index
The government should appoint a Chief Information Officer to provide strategic direction and leadership for multi-stakeholder AI adoption across government.
- ObligationApplies to: Public authorityDrafted by the Index
The government should improve procurement guidance and training so public sector bodies can effectively assess and acquire AI systems suited to their different contexts and services.
- ObligationApplies to: Public authorityDrafted by the Index
The government should ensure that training programmes equip officials with the knowledge to understand and assess how algorithmic systems make decisions before they are deployed in civil service operations.
National, cross-sector adoption 33 3 controls
- ObligationApplies to: Public authorityDrafted by the Index
Healthcare systems should integrate AI into healthcare professional training, building on existing reviews and digital academies.
- ObligationApplies to: Public authorityDrafted by the Index
Government should establish the UK as a biomedical research hub by bringing together life sciences industry and the NHS with appropriate governance to use AI for national trials.
- ObligationApplies to: Public authorityDrafted by the Index
Government should develop a plan ensuring responsible AI is suitable for addressing climate and Net Zero challenges, building awareness and incentivising developers to make accountable decisions.
Change history
No changes recorded since we began monitoring this policy.