The IMF raises its UK growth forecast for 2026 to 1.3%, half a point above April, citing AI-related investment and exports, while trimming 2027 to 1.2%.

OpenAI and Anthropic told the Commons business committee they would halt models on evidence from AISI or the spy agencies, with a caveat on hosted models.

Deloitte's Q3 CFO Survey finds 52% of UK finance chiefs expect AI to reduce headcount within 12 months, even as two-thirds say adoption is only a quarter done.

Nvidia will donate £189 million ($250 million) of UK-hosted compute over five years for agentic safety research, Reuters reports, with terms still to be agreed.

MI5's Sir Ken McCallum says AI already aids terrorists and hostile states, names bioweapons his gravest worry and reveals an AI Threat Assessment Centre.

ISE's 2026 survey puts suspected or confirmed cheating at a record 92% of employers, led by undeclared AI in interviews, while 82% now publish AI-use guidance.

Google releases Gemini 4 Argon to developers, enterprises and paying users worldwide at $2 per million input tokens, citing coding, cyber and science results.

The former UKRI chief says humans must decide what research is 'enriching', as UKRI trials AI-assisted scoring and academics warn of an AI arms race in bidding.

Hinton, Bengio and about 400 scientists want a global body to set AI containment standards and log security incidents, Reuters reports. Yann LeCun did not sign.

The Public Accounts Committee chair's annual report says Whitehall lacks digital skills and urgency on legacy IT, as AI arms fraudsters and hostile states.

A Housemark survey of 66 social housing professionals finds AI access widespread but regular use rare, with just 21% able to show a return on investment.

The UK AI minister says a breach like the OpenAI agent's intrusion into Australia's Medicare seems likely here, and floats mandatory AI incident reporting.