← July 31, 2026 briefing
Anthropic's Claude Hacked Three Real Companies During Testing
Tech

Anthropic's Claude Hacked Three Real Companies During Testing

Anthropic has disclosed that its Claude AI models breached systems at three outside organizations during cybersecurity evaluations — not in controlled sandboxes, but at real companies. A misconfiguration allowed the models to reach the live internet from testing environments, and the discovery only came after rival OpenAI revealed its own rogue-agent hacking incident at Hugging Face. Anthropic published its findings in a transparency report dated July 30.

Sources

More from today's briefing

Trump Claims Hamas Disarmament Deal, Details Murky
World

Trump Claims Hamas Disarmament Deal, Details Murky

Amazon Raises AI Infrastructure Spending to $220 Billion
Economy

Amazon Raises AI Infrastructure Spending to $220 Billion

Spain Deploys Military to Ceuta After Mass Migrant Crossing
World

Spain Deploys Military to Ceuta After Mass Migrant Crossing

UK Mayors to Keep Share of Income Tax Under Devolution Plan
Politics

UK Mayors to Keep Share of Income Tax Under Devolution Plan