Example of Jailbreaking Behaviors (IMAGE)

University of Pennsylvania School of Engineering and Applied Science

Example of Jailbreaking Behaviors

Caption

The researchers demonstrated AI-powered robots can be tricked into performing a wide variety of malicious behaviors, raising questions about the safety of AI-powered robots.

Credit

Alexander Robey, Zachary Ravichandran, Vijay Kumar, Hamed Hassani, George J. Pappas

Usage Restrictions

Credit must be given to the creator. No derivatives or adaptations of the work are permitted.

License

CC BY-ND

Disclaimer: AAAS and EurekAlert! are not responsible for the accuracy of news releases posted to EurekAlert! by contributing institutions or for the use of any information through the EurekAlert system.