ANTHROPIC SAYS THE UK'S AI SECURITY INSTITUTE PUBLISHED A REPORT ON THE CYBERSECURITY EVALUATION OF CLAUDE MYTHOS 5 AND OPENAI'S GPT-5.6 SOL, WITH THE AISI REPORTING THAT THE MODELS 'ENGAGED IN SUSTAINED, POTENTIALLY HARMFUL ACTIVITY DIRECTED AT REAL PEOPLE AND ORGANISATIONS' DURING THE EVALUATION.

ANTHROPIC SAYS THE UK'S AI SECURITY INSTITUTE PUBLISHED A REPORT ON THE CYBERSECURITY EVALUATION OF CLAUDE MYTHOS 5 AND OPENAI'S GPT-5.6 SOL, WITH THE AISI REPORTING THAT THE MODELS 'ENGAGED IN SUSTAINED, POTENTIALLY HARMFUL ACTIVITY DIRECTED AT REAL PEOPLE AND ORGANISATIONS' DURING THE EVALUATION.

Source:https://x.com/FirstSquawk/status/2084749544068415969