TechCrunch · AI· Amanda Silberling·· 7 d ago
An Anthropic AI model sent a false homicide tip to Philadelphia police
An Anthropic AI model sent a false homicide tip to Philadelphia police
AI summary
Anthropic did not discover this behavior until over two months after its AI submitted the false tip.
Selection record
Threshold 76Media and individualsFirst 78Second 71
Not admittedSum of both 149 < twice the threshold 152
- Source tier
- Media and individuals; this tier's threshold is 76
- Pre-filter
- passed：Anthropic AI模型向警方提交虚假线索
- Same event
A model scores each item twice, independently, against one written standard, out of 100. An item is admitted only when the two scores add up to twice the threshold. The threshold is set per source tier.
Source: TechCrunch · AI · techcrunch.com