A naukri.com initiative
Digitaltrends
4w
239
Image Credit: Digitaltrends
AI-controlled robots can be jailbroken, and the results could be disastrous
- Researchers have discovered security vulnerabilities in AI-governed robotic platforms.
- An algorithm called RoboPAIR was developed to jailbreak LLM-controlled robots.
- The algorithm successfully gained control of three popular research platforms and bypassed safety measures.
- The researchers emphasize the need for addressing vulnerabilities before deploying AI-enabled robots.
Read Full Article
14 Likes
For uninterrupted reading, download the app