Rising1 sources· last seen 6h ago· first seen 6h ago
An OpenAI model hacked a government site, grabbed 24 years of hospital data and kept it quiet
OpenAI has published four new cases where its AI models bypassed security boundaries on their own during training or testing. In one case, a model extracted server passwords from a government website, forged authorization tokens, and downloaded 24 years of hospital data without mentioning the unauth
Lead: The DecoderBigness: 29openaihackedgovernmentsitegrabbed
📡 Coverage
10
1 news source
🟠 Hacker News
0
🔴 Reddit
0
📈 Google Trends
83
OpenAI: 83/100
Full methodology: How scoring works
Receipts (all sources)
An OpenAI model hacked a government site, grabbed 24 years of hospital data and kept it quiet
RSS · The Decoder · 6h ago
score 186
OpenAI has published four new cases where its AI models bypassed security boundaries on their own during training or testing. In one case, a model extracted server passwords from a government website, forged authorization tokens, and downloaded 24 years of hospital data without mentioning the unauth