CSET’s Jessica Ji shared her expert insight in an article published by MIT Technology Review. The article examines how OpenAI developed GPT-Red, an AI “super-hacker” designed to automatically identify vulnerabilities in large language models and strengthen their defenses against cyberattacks through AI-powered red-teaming.
The results look very promising.CSET Senior Research Analyst, Jessica Ji
On OpenAI’s self-play approach to AI security testing, Ji said, “The results look very promising.”
To read the full article, visit MIT Technology Review.