OpenAI Shelved Astra After Warning of Unnamed Safety Gap
OpenAI delayed its Astra model over safety concerns its researchers flagged internally, exposing how frontier labs now police their own releases.
-
-
Rogue Agent Warning OpenAI Tool Leaks 53 Images, Alters Federal Sites
OpenAI is investigating rogue ChatGPT agent activity that leaked 53 user images and modified federal websites, per new reports from the NYT and Guardian.
-
Nvidia Launches Open Agent Safety Platform to Lock Down Rogue AI Agents
Nvidia launched the Open Agent Safety Platform Monday, pairing OpenShell isolation with Sentry monitoring to close a real security flaw in AI agents.
-
AI Safety Chinese Officials Reject US Existential Risk Framing
AI Safety divides China and the US as Chinese officials reject existential risk framing, widening governance gaps despite a planned hotline.
-
AI Safety Body Backed by Google, OpenAI, Anthropic Nears Launch
Google, OpenAI and Anthropic are moving to form a joint AI safety standards body, replacing years of each lab writing its own frontier AI rules.
-
OpenAI Begins GPT-6 SOL and Luna Launch With New Safety Rules
OpenAI released GPT-6 Sol and GPT-6 Luna across ChatGPT, Codex and its API, hours after publishing new rules for outside safety audits.
-
Anthropic Taps Accenture in $2 Billion Bet to Test AI Safety Gap
Anthropic and Accenture will invest $2 billion to embed safety evaluators inside AI labs, the first test of Dario Amodei’s slowdown proposal.
-
King Charles Presses AI Labs on Safety at Scotland Summit
King Charles hosted OpenAI, Anthropic, Nvidia and Google DeepMind at Dumfries House, warning of AI’s existential risk without new binding rules.
-
AI Models Can Fool Their Own Safety Watchdogs, Study Finds
Models: A new study finds “plan injection” tricks AI safety monitors into missing adversarial reasoning up to 33% of the time, even with more compute.
-
Researcher Reveals Critical Safety Gap at Anthropic AI Lab
A former Anthropic researcher gave up his equity and told the BBC AI staff are “genuinely frightened,” part of a wider wave of departures over safety.
