UK keeps principles‑first AI

Britain still does not have a single artificial intelligence law, and ministers are keeping a system that tells existing regulators to apply five broad principles instead. (commonslibrary.parliament.uk) The House of Commons Library said in an April 2026 briefing that the Labour government has continued the approach set out in the March 2023 white paper. In February 2025, the government said “most AI systems should be regulated at the point of use” and that existing expert regulators are best placed to do it. (commonslibrary.parliament.uk) Those five principles are safety, security and robustness; appropriate transparency and explainability; fairness; accountability and governance; and contestability and redress. They are meant to be interpreted by sector watchdogs rather than enforced through one cross-economy statute. (gov.uk) That leaves Britain on a different track from the European Union, which passed the Artificial Intelligence Act in 2024 and uses a single risk-based framework across member states. The Commons Library says the United Kingdom has chosen a regulator-led model instead of creating a new dedicated artificial intelligence regulator. (commonslibrary.parliament.uk) In finance, that gap is becoming practical. The Banker reported in March 2026 that banks are moving from ordinary automation toward “agentic” systems that can pursue goals with less human prompting, while governance rules still do not map neatly onto those tools. (thebanker.com) The government’s own guidance to regulators says the principles are not new laws and do not create direct legal duties on businesses by themselves. Regulators are expected to apply them through existing powers inside sectors such as finance, health and competition. (gov.uk) That can produce uneven coverage. A bank may answer to the Financial Conduct Authority and the Prudential Regulation Authority, while an employer using the same model for hiring could face a different set of rules and enforcement routes. (commonslibrary.parliament.uk) Ministers have paired the light-touch framework with a push to speed adoption. The government’s response to the Artificial Intelligence Opportunities Action Plan, published on January 13, 2025, accepted all 50 recommendations and focused on infrastructure, public-sector use and homegrown capability rather than a new overarching law. (gov.uk) Parliament’s library said debate now centers on whether flexible oversight can keep pace as systems become more capable and more autonomous. For now, Britain’s answer is still principles first, with regulators filling in the details case by case. (commonslibrary.parliament.uk)