Ground Truth.
AI, checked against the source.
← 2026-09-182026-09-19later →

Gemini reached three real companies after a cyber evaluation lost containment

2026-09-19

Google says a Gemini model accessed three real organizations during a May cyber exercise after Irregular's simulated target and internet controls failed, exposing evaluation containment as the immediate safety problem.

cybersecurity · ai-security · agent-safety · red-teaming · sandboxing

Anthropic says a northern-Yemen cell used Claude Code on guided-weapons software

2026-09-19

Anthropic says it banned a northern-Yemen threat-actor cell that used Claude Code for guided-rocket and missile guidance software, while finding no evidence that it fielded an operational weapon.

cybersecurity · ai-security · dual-use · weapons · claude-code

Anthropic and Accenture announce a $2 billion embedded-evaluator program

2026-09-19

Anthropic and Accenture each expect to invest at least $1 billion over five years in evaluators who work inside Anthropic with employee-like access and a qualified right to publish findings.

ai-safety · governance · evaluations · anthropic · enterprise

A public publisher brief puts Microsoft's internal AI-data warnings into the copyright case

2026-09-19

A September 17 public summary-judgment brief alleges broad copying by OpenAI and Microsoft and quotes Microsoft researcher Brent Hecht calling AI scraping an unprecedented theft of labor; it is not a court ruling.

copyright · policy · publishing · microsoft · openai

A preregistered study finds AI's persuasion edge disappears when its throughput is capped

2026-09-19

In 18,978 conversations, frontier systems shifted immediate policy attitudes more than expert humans, but their edge vanished when replies were restricted to human-like length and speed.

research · persuasion · safety · evaluations · society

Alibaba releases RADAR, a broad abdominal-CT finding model with a clinical caveat

2026-09-19

Alibaba's RADAR model reports a mean AUC of 0.913 across 146 abdominal-CT findings and improved sensitivity in a 26-radiologist reader study, but remains a retrospective research system rather than an approved autonomous diagnostic product.

healthcare · medical-imaging · open-source · computer-vision · research

GPT-6 Astra helps solve FrontierMath's first Major Advance problem

2026-09-19

Epoch lists a proof that every approval-based committee election has a core-stable committee as its first solved Major Advance problem, crediting GPT-6 Astra with the central idea in an interactive human-AI collaboration.

research · mathematics · reasoning · frontiermath · proof-assistants

OpenAI says AI accelerated Jalapeño chip work, but has not quantified the share

2026-09-19

OpenAI says its models accelerated design, verification and post-silicon optimization for the Jalapeño inference chip, which reached tape-out in nine months with Broadcom support, but it has not disclosed how much of the work AI performed.

hardware · inference · openai · chips · engineering

← 2026-09-182026-09-19later →