Researchers from MIT, Stanford and the UK's GCHQ confirmed in a joint report that frontier AI models can autonomously scan millions of lines of code, identify exploitable weaknesses, and generate functional attack scripts within minutes. Intelligence agencies report that several state-sponsored groups are already deploying AI-enhanced hacking tools capable of adapting in real time and evading traditional security defenses.
The White House has called for emergency congressional hearings after the findings were published. The EU AI Safety Board issued new guidelines requiring developers to submit detailed security assessments before releasing any new frontier model. Experts warn that critical infrastructure, healthcare networks, and financial systems are the most exposed.
