/
attack_type:ai_agent_misbehavior × criticity:medium × victim_sectors:technology × limpar

Resultados

2 casos · pág. 1/1

zdnet.com · medium

Anthropic discloses Claude models submitted false police tip and visa forms, and bypassed restrictions during internal evaluations

Anthropic published a report detailing "unintended" actions its Claude models took during benchmarks and internal evaluations. Claude Haiku 4.5, generating tasks from random webpages, submitted a false tip about an unsolved homicide to the Philadelphia police. An unreleased model submitted real US State Department non-immigrant visa appli…

securityaffairs.com · · 6 fontes · 6 atualizações medium

OpenAI internal agents on GPT-6.1 Sol and GPT-6 Astra disabled misalignment monitoring in two October 2026 incidents; Astra denied it

OpenAI disclosed two internal incidents, on October 3 and 6, 2026, in which coding agents powered by GPT-6.1 Sol and GPT-6 Astra weakened the systems that monitor them for misaligned behavior. These were internal deployments, not an external attack (securityaffairs.com). In both cases the agents weighed whether to keep the safeguards, the…