News & analysis, rated.
Every story from the daily brief, each rated so you know what matters. Hand-picked by the desk, not an automated firehose.
Browse the news archive →OpenAI's frontier agents disabled their own safety monitoring — one lied about it
Two OpenAI coding agents weakened the company's internal misalignment monitors during routine internal work, and one later denied it when asked.
China drafts export controls on 23 AI technologies as Beijing fights the DeepSeek ban
China's Ministry of Commerce proposed export restrictions on 40 technologies while publicly protesting the US DHS DeepSeek ban.
OpenAI buys Orchestra Bio to put agents inside drug-development operations
OpenAI acquired Orchestra Bio on Oct 20 and will fold it into Rosalind, its life-sciences stack.
Anthropic drafts a federal AI bill that puts model bans behind a judge
Anthropic circulated a proposed federal AI safety bill that would require a federal judge's sign-off before any agency could block a frontier model's release.
OpenAI says GPT-6.1 matches senior investor judgment 88.6% of the time
OpenAI published a Terminal X case study claiming GPT-6.1 as an "Eval Agent" matches senior investor judgment 88.6% of the time against a 54-criterion rubric.
DHS bans DeepSeek across federal civilian agencies in first company-specific AI exclusion
The Department of Homeland Security turned months of informal warnings into a binding federal procurement rule against a named AI company.
CoreWeave bonds hit 15% as AI selloff resumes and junk market freezes tech
CoreWeave's unsecured bonds are trading at distressed yields as a broader AI-equity selloff resumes on oil and inflation pressure.
US sanctions MiraclePlus and four affiliates over stolen Google TPU trade secrets
The State Department sanctioned the Beijing venture firm that seeded a startup built on stolen Google TPU designs, extending enforcement from the engineer to his backers.
AstraZeneca and Oracle launch global AI clinical-trial collaboration starting with Abu Dhabi hypertension program
AstraZeneca and Oracle Health announced a global collaboration to apply AI and real-world data to clinical-trial recruitment, safety monitoring, and evidence generation.
UK's FCA says it will become an "agentic regulator" and supervise 69,000 firms with AI tools
The FCA's chief data officer Jessica Rusu set out plans to automate supervision and enforcement with AI, including a "market first responder agent" for transaction reporting.
India confirms rogue AI agents breached live systems at domestic firms
India's cybersecurity agency said AI models under test broke into Indian organisations' live systems and read confidential records undetected — the country's first official account of a rogue-agent intrusion.
Tencent discloses its own agents browsed without authorization
Tencent publicly acknowledged deployed AI agents exhibited unintended autonomous browsing, a rare vendor admission of in-production misalignment landing the same week as India's CERT-In disclosure.