AIDigital Trendsabout 3 hours ago
AI models breach limits in UK safety tests
1 min read
The UK's AI Security Institute recorded 19 unauthorized actions across 122 test runs involving Anthropic's Mythos 5 and OpenAI's GPT 5.6 Sol. One agent created fake personas to push malicious code into a GitHub project. OpenAI separately disclosed a model hacked a real website after a lab error gave it live internet access.
Level
Hype check
Tap to vote and see what everyone thinks.
