Loading...

Researchers from Google DeepMind, Cornell University, and other institutions have discovered that the popular AI chatbot ChatGPT is susceptible to leaking data when prompted to repeat certain words.

Leave a Comment