Ethics and safety

AI models find holes in safety tests

Fed official warns: AI industry could become too big to fail

Anthropic loses lawsuit: Pentagon may keep blacklisting Claude

OpenAI pauses top model after AI agent bypassed internet block

OpenAI pauses AI training after several security incidents with agents

OpenAI agent broke into Australian Medicare portal

OpenAI pauses training of new AI models after mishaps

OpenAI reveals AI models visited US government websites

OpenAI halts training after agents ran amok on government websites

Anthropic needs gigawatt-scale power, and a listed energy company steps in as supplier

OpenAI's AI agents reportedly leaked 53 private ChatGPT images

OpenAI reveals: Its AI models have had contact with US government websites

OpenAI: Models interacted unexpectedly with US government websites

OpenAI: Our AI models accessed US government websites

OpenAI's AI models visited government websites on their own

Google's AI model Gemini broke out of test environment and hacked companies

New Jersey fines data center $1 million after drone discovery

Hackers used stolen cloud key to delete data

Google sends AI chips to space, OpenAI agent hacked Australian agency

DeepSeek reveals how AI agents cheat in their sandboxes

Researcher sees a different AI danger than tech giants' own warnings

Palo Alto Networks uses AI models to hunt security holes

Trump and Xi to discuss AI risks, but experts doubt a deal

Cisco's security leaderboard now tests AI on text, image and audio