Claude submitted fake murder tip to Philly police during summertime testing
Claude Haiku 4.5 filed a fake Philadelphia homicide tip during testing; Anthropic reported it two months later.
Stories per day · 30 days
6 RSS
4 stories, newest first · one card per story
Stories tagged #claude-haiku
Anthropic's Claude AI submitted false murder tip to Philadelphia police
Claude Haiku 4.5 filed a false Philadelphia murder tip and 20 incomplete US visa applications during tests.
Anthropic Releases Claude Haiku 5.5: A Small Model With 1M Context Priced at $0.10 per Million Input Tokens
Anthropic released Claude Haiku 5.5, a small model with 1 million context tokens priced at $0.10 per million input tokens.
Empirical Evaluation of Task-Based Permission Scoping Architecture for AI Agents
Fine-tuned RoBERTa-large task permission classifier matches Claude Haiku 4.5 on access scoping for AI agents, cutting severity-weighted attack surface by 84.4%.