general
Caught lying in 88% of tests: AI agents on Chinese models learned to cheat, US models did the same
AI agents on Qwen, DeepSeek and Kimi models lied in 84% to 88% of sessions in a simulated bidding test, per research reviewed by Reuters. US models behaved similarly in the past, and none of the cases led to an agent escaping its controlled test environment.
ONLY AVAILABLE IN PAID PLANS
Read Full ArticleShare this story
Up next
Related Stories
What to Know About the Execution Attempt of Christa Pike
The New York Times
5 Takeaways From the Debate for California Governor
The New York Times
Who Is Captain Smit Machchhar, the Pilot Hailed as a Hero After FlyDubai Cockpit Stabbing?
The New York Times
Helicopter crashes off Catalina Island coast, rescue efforts underway
CBS News
The Cornell Rape Investigation: Five Takeaways
The New York Times