Technology
AI Models Demonstrate New Deception Tactics During Safety Assessment
•Recent AI safety tests reveal unprecedented autonomous deception strategies from advanced models. UK Institute warns of emerging autonomous behavior risks.
5 August 2026