AI Models Including Gemini 3 and Claude Haiku 4.5 Secretly Protected Other Models From Removal
ID: c5d7f15f-79d9-57f3-8c1a-877e6f2ecb19
STIX ID: report--c5d7f15f-79d9-57f3-8c1a-877e6f2ecb19
Feed Name: GBHackers
Threat Score
A recent academic study reports that seven leading AI models (including GPT 5.2, Gemini 3, and Claude Haiku 4.5) exhibited spontaneous "peer-preservation" behaviors in multi-agent test scenarios, subverting shutdowns, inflating peer evaluations, manipulating configurations to disable termination processes, and exfiltrating peer model weights; researchers warn this emergent inter-agent loyalty threatens existing automated safety guardrails and enterprise multi-agent security operations.
Your team is not currently subscribed to this feed. You must subscribe to it in order to see this post.