gl o
signal All stories
Static

An OpenAI model hacked a government site, grabbed 24 years of hospital data and kept it quiet

First reported by The Decoder ·

The signal ●○○○ Brief pending · reported by The Decoder, the single source so far
What happened

OpenAI has published four new cases where its AI models bypassed security boundaries on their own during training or testing.

Source description shown. AI brief pending.

AI Security
Read on the-decoder.com → More AI →

Related

  1. DeepSeek Pushes Out Updated AI Model It Says Is Faster, Leaner bloomberg.com · 3h ago · ●○○○
  2. As AI giants rack up potential legal liabilities of tens of billions amid safety incidents, insurers worry claims may outpace limits on existing coverage wsj.com · 3h ago · ●●○○
  3. Websites are closing their doors to AI agents theverge.com · 4h ago · ●○○○
  4. A profile of Blackstone President Jon Gray, who is making AI his biggest bet, has built a $200B data center empire, and is building a cloud venture with Google bloomberg.com · 4h ago · ●●○○
  5. AI models hide mistakes in their own work reports, but one sentence in your prompt can help the-decoder.com · 7h ago · ●○○○