AI regulation in United Kingdom (2026)

DSIT published the first annual progress report showing 38 of 50 action plan commitments met within 12 months, including the Isambard-AI supercomputer at Bristol, five designated AI Growth Zones, and one million AI training courses ahead of schedule. Progress on regulatory reform lagged, with half the regulation recommendations still outstanding.

The UK's first primary legislation directly addressing AI became law; Part 7 requires the government to publish within nine months an economic impact assessment on AI and copyright and a report on the use of copyright works in AI training, directly responding to a major tension between rights-holders and AI developers.

Lord Holmes of Richmond reintroduced the private member's AI Regulation Bill, proposing a new statutory AI Authority to coordinate sector regulators, close governance gaps, and accredit AI auditors. The bill lacked government backing and faced limited parliamentary time, but amplified pressure for dedicated AI legislation.

DSIT renamed the AI Safety Institute to the AI Security Institute, shifting focus from broad safety concerns (bias, free speech) to security-critical harms — AI-enabled cyberattacks, CBRN weapons development, and child sexual abuse imagery. A new criminal misuse team was launched in partnership with the Home Office.

The government endorsed all 50 recommendations by tech entrepreneur Matt Clifford to position the UK as a global AI leader, covering AI Growth Zones, expanded supercomputing, AI adoption across public services, and homegrown AI development — projecting £400 billion in additional GDP by 2030.

The UK co-hosted the second global AI safety summit with South Korea; 28 nations signed the Seoul Declaration and Ministerial Statement, which for the first time established shared risk thresholds for when frontier AI capabilities could pose 'severe risks' without adequate mitigations — advancing the Bletchley framework into concrete commitments.

DSIT confirmed the UK would not introduce AI-specific primary legislation, maintaining its decentralised approach while committing regulators to mandatory annual horizon-scanning and joint cross-regulator AI workplans. The response closed the consultation opened with the March 2023 White Paper.

The UK hosted the first global AI Safety Summit at Bletchley Park; 28 countries and the EU signed the Bletchley Declaration committing to cooperative frontier AI testing before deployment. PM Sunak simultaneously launched the AI Safety Institute — the first state body anywhere dedicated to evaluating advanced AI model risks — evolving from the Frontier AI Taskforce.

The Prime Minister created the Frontier AI Taskforce, backed by £100 million and chaired by Ian Hogarth, as the first unit inside any G7 government dedicated to evaluating frontier AI model risks. It conducted the UK's first government-level frontier AI evaluations before evolving into the AI Safety Institute in November 2023.

DSIT published the UK's defining AI regulatory framework document, explicitly rejecting an EU-style horizontal AI Act in favour of five cross-sector principles (safety, transparency, fairness, accountability, contestability) applied by existing sector regulators with no new AI-specific legislation. This remains the cornerstone of the UK's regulatory architecture.

The government published its first National AI Strategy across three pillars — long-term ecosystem investment, AI-enabled economic transition, and national and international AI governance — setting the ambition to rank the UK alongside the US and China as a global AI superpower and providing the political mandate for all subsequent regulation.

As part of the Industrial Strategy, the government announced the AI Sector Deal — a package of up to £950 million in government, industry, and academic funding — including 1,000 new AI PhD places and commitments on compute and data access. It translated the 2017 Hall-Pesenti recommendations into funded policy.

The Lords Select Committee on Artificial Intelligence published a landmark nine-month inquiry report drawing on 280 witnesses, recommending an AI ethics code, a cross-regulator AI governance body, and stronger data access for SMEs. It was the UK's first major parliamentary scrutiny of AI and shaped the government's ethical AI discourse for years.

Government-commissioned independent review by Dame Wendy Hall and Jérôme Pesenti made 18 recommendations on AI skills, data access, research, and commercialisation, and proposed the Alan Turing Institute as the national AI hub. It established the UK's foundational AI industrial policy and directly led to the 2018 Sector Deal.