The middle category is the one that changed the most in 2026.
An agent making a wrong choice used to mean a bad chat. Now it can mean a bad transaction, a leaked file, or an unrecoverable action.
Serious reading on all three, from the International AI Safety Report
👇
https://internationalaisafetyreport.org/publication/international-ai-safety-report-2026
An agent making a wrong choice used to mean a bad chat. Now it can mean a bad transaction, a leaked file, or an unrecoverable action.
Serious reading on all three, from the International AI Safety Report
👇
https://internationalaisafetyreport.org/publication/international-ai-safety-report-2026