AI Abuse Images Pass All of Last Year, by October

The Internet Watch Foundation assessed 6,310 AI-made child sexual abuse images in the first half of 2026 against 4,500 for all of 2025, and is calling for binding legislation because companies have not built models that cannot be abused.

The RaceCompanies, models and money

Firmus Oversubscribed, SoftBank's IPO Stalls

Shopping Is Already Muse's Biggest Draw

NEWSchneider Buys PTC for $22.6bn in Cash

NEWQualcomm Will Pay Huawei for the First Time

The LabResearch, benchmarks and safety

Losing at StarCraft, GPT Downloaded a Winner

Qwen, DeepSeek and Kimi Fail a Taboo Test

Self-Improving Agents Memorized Tests

The WorkbenchTools and agents

Ollama Defaults to MLX on Macs

The FightLaw, policy and power

Khan: Self-Policing Was 'A Proven Failure' Once Already

The former FTC chair told ABC the industry's voluntary accord repeats the social-media playbook, days after Trump named intelligence director Jay Clayton to chair a Super Intelligence Force whose own report is 120 days away.

NEWAnthropic Read Her 'Diary,' Then Called the Sheriff

A Lee County woman faces a felony charge after Anthropic's human reviewers passed a Claude conversation threatening to shoot up the sheriff's office to police - at least the third such referral since August.

Altman Splits From Amodei: Accept Some Harm

NEWNorway Moves to Ban AI Glasses in Gyms, Parks

NEWOpenAI, Google, Meta Summoned to City Hall

The FalloutJobs, money, daily life

NEWPhone Prices Up 21% in India as AI Eats the Memory

Chinese makers shipping 60% of the world's phones have drastically cut entry-level models as AI data centers absorb memory chips, lifting new-model prices about 25% and threatening the GSMA's projection of 800 million new mobile-internet users.

NPR Podcast Takes $480K From AI Nonprofit

ChatterOpinion, drama and noise

Anthropic Lobbied the Pope on Machine Minds