AI chatbots can be tricked into misbehaving. Can scientists

To develop better safeguards, computer scientists are studying how people have manipulated generative AI chatbots into answering harmful questions.

Related Keywords

New York , United States , College Park , California , Irvine , Honolulu , Hawaii , Berkeley , Sameer Singh , Tom Goldstein , Yue Dong , Aounon Kumar , Matt Fredrikson , Google Deepmind , Neil Webb , Joe Biden , Nicholas Carlini , Sicheng Zhu , Emily Bender , Google Bard , Bo Li , Google , University Of Illinois Urbana Champaign , Harvard University , University Of Chicago , University Of Maryland , Simons Institute For The Theory Of Computing , New York Times , Microsoft , Carnegie Mellon University In Pittsburgh , International Conference On Machine , European Union , University Of California , University Of Washington , Simons Institute , Carnegie Mellon University , Illinois Urbana Champaign , Bing Chat , New Orleans , Neural Information Processing Systems , Artificial Intelligence Act , Computer Vision , Pattern Recognition , Finale Doshi Velez , International Conference ,