U.K., Nov. 18 -- UK Parliament issued the following news:
Britain risks undermining AI's potential £400 billion economic prize by leaving critical safety decisions to the companies building the most powerful systems, MPs warn today.
Read the report [PDF]
Following hearings with AI researchers, former industry employees and senior representatives of OpenAI, Google DeepMind, Anthropic and Meta, the Business, Innovation, Science and Trade Committee identifies ten gaps in UK safeguards that leave the public exposed and businesses uncertain about how risks will be controlled.
In its report, Guardrails for Growth: Making AI Safe, the Committee draws on OECD productivity estimates to calculate that AI could leave Britain's economy up to £400 billion larger after ten years - while the UK's internationally recognised expertise in AI testing and evaluation could create a world-beating new export industry in AI assurance.
But the Committee warns effective safeguards must keep pace, to protect the public and give more businesses, especially SMEs and micro-businesses across the country, confidence to adopt the technology, and support for consumers to use the technology safely and responsibly.
The Committee's hearings exposed weaknesses in independent testing, incident reporting and powers to intervene. Former employees questioned the industry's safety culture and described commercial pressures, contractual restrictions and financial disincentives that can discourage staff from raising concerns.
The AI Security Institute (AISI) is an internationally recognised centre of expertise in AI testing, research and evaluation, but without a legislative framework for frontier AI safety and governance its invaluable work depends on voluntary cooperation and its ability to access talent.
Developers told MPs they would not release unsafe models, but the evidence showed that companies apply different safety thresholds and remain responsible for judging whether their own systems meet them.
The Committee heard that independent evaluators lack guaranteed access, reporting arrangements remain fragmented, and some models cannot reliably be recalled once widely copied and distributed. The Committee concludes that AISI's internationally respected technical expertise must be complemented by enforceable legal safeguards and a separate statutory regulator.
The Committee calls for an AI Safety and Security Bill in the next parliamentary session as part of a clear legislative roadmap that should combine targeted action, duties to prevent harm, a clear liability regime, criminalising misuse of AI for harm and an international convention to enshrine global rules of the game.
It also recommends a new regulator is appointed - distinct from the AI Security Institute - with legal powers to require independent verification of safety claims, pause high-risk training and testing, delay or prohibit deployment, and suspend services where risks cannot be controlled. New safeguards must kick in while systems are being developed.
The report identifies ten critical weaknesses that ministers must address:
Missing definitions of unacceptable harms and risk thresholds.
Lack of statutory powers to access models.
Lack of independent verification of developers' safety claims.
Insufficient controls on training, testing, internal deployment and containment.
Absence of mandatory, independently scrutinised safety assessments before high-risk systems are deployed.
Absence of mandatory reporting of incidents and near misses.
Lack of powers to restrict or halt unsafe development and deployment.
Insufficient safeguards for models whose release cannot be reversed.
Inadequate liability arrangements and routes to redress.
Insufficient protection for whistleblowers.
The Committee says ministers, accountable to Parliament, must determine what risks society should accept. AISI should provide technical evidence to inform those decisions, while the regulator must have clear powers to enforce the resulting requirements.
Firms, the public and Parliament are entitled to know who sets the limits, on what evidence, with what right of challenge, and who acts when those limits are breached.
Developers of designated high-risk systems should be required to substantiate their safety claims and provide independent evaluators with meaningful access. Material changes in a system's capabilities, safeguards or use should trigger reassessment, with restrictions or suspension where risks cannot be adequately controlled.
The Government must introduce mandatory incident reporting and protected channels for whistleblowers. It must also establish effective protections and routes to redress for individuals, clarify responsibility across AI supply chains, and ensure appropriate criminal offences are available to deter harmful misuse and activities that enable it.
International cooperation must reinforce domestic action. Within three months, ministers should publish objectives for the UK's 2027 G20 presidency, identifying negotiating priorities, responsible ministers and milestones.
These should include cooperation on model access, safety assessments, incident reporting and high-risk releases, alongside an assessment of options for a multilateral convention or treaty. The pursuit of international agreement must not delay safeguards at home.
Chair comment
Rt Hon Liam Byrne MP, Chair of the Business, Innovation, Science and Trade Committee, said:
"Britain has a huge prize within reach: AI could expand our economy by £400 billion larger in just a decade. We have world-leading expertise in AI testing and the chance to create a new world-beating export industry in AI assurance. But clear rules and credible safeguards are now needed to help British businesses invest, innovate and capture those gains.
"We cannot give today's AI leaders unfettered freedom to behave like modern-day Dr Frankensteins, creating dangers beyond anyone's control. Our hearings exposed gaps in independent scrutiny, incident reporting and powers to stop unsafe systems. Once some models are released, they cannot reliably be recalled. Protection must begin while systems are being built and tested.
"AI firms must be required to substantiate their safety claims, independent experts must be able to scrutinise them, and a regulator must have the power to intervene when risks cannot be controlled. Ministers, accountable to Parliament, must decide what risks the public should be asked to bear.
"Britain is world recognised for our history in creating fair rules of the game. Now we need to use Britain's G20 presidency to secure international cooperation alongside enforceable safeguards at home. That is how we protect the public and give businesses the confidence to seize this extraordinary economic prize."
This is the first report in the Committee's long-term inquiry into AI adoption, safety and governance. It focuses primarily on domestic measures required in the short and medium term.
Alongside the report, the Committee is launching a call for evidence on the UK's role in securing a global convention on AI safety and security through its 2027 G20 presidency.
Further information
About Parliament: Select committees
Visiting Parliament: Watch committees
Business Innovation, Science and Trade Committee
Image: Adobe stock
Disclaimer: Curated by HT Syndication.