Posted 2/17/2024, 9:00:00 AM
OpenAI Uncovers State-Sponsored Groups Abusing AI Services, Proposes New Security Measures
- Foreign state actors used OpenAI services to train operatives, prompting calls for new AI security infrastructure
- OpenAI identified 5 "malicious" state-affiliated groups abusing its services from China, Iran, North Korea, and Russia
- OpenAI proposed new monitoring and disruption tech to catch bad actors, along with cooperation across AI platforms
- Expert argues new infrastructure and regulations, like in banking, are needed to effectively combat AI misuse
- Microsoft and OpenAI pledge new notification system and countermeasures, but expert remains skeptical without government help