Anthropic's Mythos found flaws in classified US systems
Anthropic's Mythos model uncovered vulnerabilities in classified US government systems in hours during a restricted testing programme, the AP reports.
AI News — Page 42
The latest AI developments shaping UK business and technology
British Army AI cuts war-planning cycle from 72 hours to one
The Army's ASGARD system has compressed corps-level planning from 72 hours to one, its top general says, part of a £1bn battlefield AI push.
Met chief warns police risk being outwitted by AI crime
Sir Mark Rowley says criminals exploiting AI move faster than forces bound by slow procurement and calls for fewer, larger police forces.
Tech stocks slide as investors question AI spending
A global sell-off hit chipmakers and AI stocks on Tuesday, reviving dot-com comparisons as seven firms make up 30% of the S&P 500's value.
UK shoppers want AI buying help but distrust paid results
A Commerce and PayPal survey finds 64% of UK consumers keen on agentic shopping, yet 46% reject letting brands pay for preference in AI results.
Welsh college sets out AI pledge for staff and learners
Coleg Cambria has launched an AI Pledge built on five principles, among the first such frameworks in Welsh further education, keeping humans in the loop.
UK backs Oxford and UCL labs to make AI cheaper to run
Two government-backed labs will share up to £60m to build open-source AI that runs on ordinary hardware, part of a £1.6bn UKRI strategy.
AI law firm wins English court case in a legal first
Garfield AI prepared every document for a £7,000 county court claim won by a freelancer who paid about £400 — believed to be the first trial won using an AI lawyer.
IBM taps OpenAI's 'Daybreak' models for enterprise security
IBM has joined OpenAI's Daybreak Cyber Partner Program, launching an AI application-security service to find and validate software vulnerabilities faster.