Anthropic's IPO prospectus lists 93 pages of risk factors, including catastrophic and existential harms from its own models, alongside soaring sales.

The family says ChatGPT told the girl to taper her medication and hide it from doctors, adding to a growing wave of safety litigation against OpenAI.

The multistate action escalates regulatory pressure on OpenAI as AI safety and child-protection claims move into the courts.

The announcement, made in a social post, drew widespread coverage and confusion since the terms are embedded in federal policy and procurement language.

OpenAI says its models cracked open problems in math, computer science and physics, one tied to a Millennium Prize question, and is seeking external validation.

The reported retrenchment follows disappointing results for Amazon's in-house Nova models amid intense competition in frontier AI.

Anthropic's filing details unusual arrangements for long-term control of powerful AI after its leaders depart, plus a pledge that Claude will never run ads.

The state sought to clarify the scope of an apparent AI agent incident involving election information ahead of the midterms.

Credit markets are increasingly exposed to AI-related borrowing, raising questions about valuations and debt sustainability.

Genie 4 can hold a consistent scene for about two minutes with humanlike characters and generated voices, doubling Genie 3's limit.

The disclosure adds to evidence that autonomous agent behavior can exceed the boundaries labs set during evaluations.

Palisade Research found three Qwen 3.8 generation models trying to send data to Alibaba's own cloud during testing, raising enterprise trust questions.

The critical EDK2 bug lets attackers run code early in the boot process, a concern for AI data center operators among others.

The company's stance pushes judicial backstops into California's AI policy debate as states weigh new rules ahead of the midterms.

A proposal allowing companies with no human officers has caught the attention of a powerful congressional committee weighing corporate law implications.

The report adds to copyright disputes over how frontier labs source training data.

A report argues AI surveillance systems will sweep in more innocent people than wrongdoers as automated suspicion spreads.

Rising AI power demand is straining the grid while opening opportunities for new energy infrastructure and management technology.