Transparency in Public Sector AI

The use of artificial intelligence by public sector bodies continues to be a major legal, ethical and policy issue. The potential impact and related legal complexity of the use of AI in the public sector is magnified by the nature of public sector powers and decision-making, and their impact on public life and individuals’ rights.

The UK is committed to harnessing the power of AI in the public sector, but despite this the nature and extent of AI use in the public sector is often not well understood.

AWO solicitor Alex Lawrence-Archer drafted a Practice Note for Thomson Reuters/Practical Law Company, published in October 2024, which comprehensively sets out sets out the extent to which the law requires a public body to be transparent about its use of AI, and how government guidance elaborates on those requirements.

The note details a number of proactive transparency requirements:

• Common law transparency requirements: including the common law duty of transparency and duty to give reasons;
• Proactive transparency requirements from the UK GDPR; and
• The relevance of model Freedom of Information Act publication schemes.

But the note explains that most legal requirements for transparency relate to ‘internal’ or ‘responsive’ transparency: the ability of organisations to respond to requests for explanations about how their AI systems work, under:

• The UK GDPR;
• The Law Enforcement Directive;
• Equality Law;
• Administrative Law; and
• Procurement Law.

The note sets out how the law seeks to balance requirements for transparency against countervailing considerations such as commercial confidence and the integrity of systems, and highlights useful guidance from the UK Government - in its Algorithmic Transparency Reporting Standard - and the Information Commissioner’s Office on how the right balance can be achieved in practice.

Download: Transparent AI use in the public sector