BoE's Bailey: Test AI Before Regulation — SkimNews
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- Andrew Bailey, Governor of the Bank of England, wrote in his first-ever Substack post that regulating AI 'is not the right place to start' — calling first for 'rigorous' testing to find vulnerabilities rather than halting development, which he said would forgo 'immense' benefits.
- Bailey endorsed the UK's AI Security Institute as part of forming standards but urged 'the pace of progress must accelerate,' arguing that unexpected model behavior 'is not evidence that testing has failed, rather it is evidence of why testing is necessary.'
- President Trump rejected industry calls to slow AI development, saying the US leads China in the AI race because 'whoever wins AI, wins.'
- Trump hosted a summit with OpenAI and Anthropic where executives signed a 'morally binding' document he described as a 'form of protection' from AI risks, placing responsibility for safety on the companies themselves.
- OpenAI recently announced it would not release its latest AI model due to safety concerns, as leaders of firms including Anthropic push for an internationally coordinated approach to AI risk assessment.
Why it matters: A G7 central bank governor is publicly endorsing a 'test before regulating' framework for AI, lending institutional weight to the UK AI Security Institute's voluntary testing model and signaling that any binding AI rules remain distant while safety testing carries the immediate load.
Ask SkimNews
