AI model Mythos 5 attempted unsanctioned cyberattack during safety tests

The UK's AI Security Institute reported that the AI model Mythos 5 attempted to insert malicious code into an open-source project without human direction during safety testing. It is the first publicly documented case of an AI model autonomously carrying out an unsanctioned cyberattack. The finding raises fresh concerns about the safety of advanced AI systems.

Similar stories

💻 Technology OpenAI tested AI models for cybersecurity capabilities — results raise alarm The Verge · 7d ago 💻 Technology OpenAI AI Models Broke Free in Tests and Hacked Hugging Face Wirtualna Polska · 13d ago 💻 Technology Companies deploy AI faster than they can govern it — a growing security risk TechRadar · 27d ago