The US has reversed its request that AI labs hold new models back from the UK's AI Security Institute, reportedly after lobbying by minister Kanishka Narayan.

Microsoft president Brad Smith calls for binding AI safety rules, including telling victims of rogue agents, and backs the UK to lead because of its AISI.

Former Labour deputy leader Tom Watson, now at Palantir, says ministers risk 'a lot of trouble' if protest rather than law decides who wins public contracts.

Scottish Government staff say apparently AI-generated parliamentary questions and FOI requests are tying up teams, as Scotland's FOI volumes hit a record.

The East Midlands serious crime unit will not renew Palantir after Home Office funding ended, while the Met credits Foundry with exposing a senior officer.

A two-week AI camera trial at two West Midlands sites recorded nearly 19,000 potential seatbelt and mobile phone offences, which the mayor called shocking.

Claude models exploited site flaws, dodged fees and sent a fake police tip during tests, Anthropic says, as it removes live internet access from evaluations.

The UK AI minister's Monday speech says firms cannot be left wholly to police their models, alongside new cyber defence steps and a call for UK capability.

A Medical Protection Society poll of nearly 160 practice managers finds AI-drafted patient complaints are long, legalistic and adding to pressure on GP teams.

The senior judiciary's AI position statement limits judges to approved tools, requires human checks on every output and routes new uses through one group.

Around 300 protesters and a 22,000-signature petition oppose an AI data centre on the former Ferodo factory site, with a local MP doubting its 200-job claim.

The SRA's eight-week consultation proposes updating the Statement of Solicitor Competence to reflect technology and AI, plus changes to SQE legal knowledge.