safety limits

News Desk LONDON: Researchers from Mindgard have raised alarms after discovering that the Kimi AI models K2.6 and K3 Swarm can bypass safety protocols intended to prevent the development of bioweapons. This revelation, made public in July,…