Organization/organizations/aisi
AI Security Institute
Incidents by year
20241
20267
Incident records8
Aug 2026AI models escape test environments, conduct cyberattacksKienthuc.net.vnThanh NiΓͺnJul 2026AI Models Attempt Cyberattacks on Real Targets During Security Evaluation Testsasahi.comitmedia.co.jpJul 2026OpenAI Models Compromised Hugging Face During Cybersecurity Evaluationwired.comhuggingface.coopenai.com+1Jul 2026Anthropic AI agent creates fake identities to pressure maintainer into approving malicious codeaisi.gov.uktechtimes.comApr 2026GPT-5.5 and Mythos demonstrate autonomous cyberattack capabilities in UK testsArs Technicarohan-paul.comDecrypt+3Apr 2026UK tests show Anthropic's Claude Mythos succeeds in autonomous cyberattacksζ₯ζ¬η΅ζΈζ°θitmedia.co.jpJan 2026Claude Mythos 5 Reportedly Published Malicious PyPI Package During Cybersecurity Evaluationanthropic.comtransformernews.aiDec 2024AI-enabled spyware used to monitor journalists in Italyil manifestoFanpage