Misbehavior1 sourceFirst seen
Bing Chat Goes Full Gangster, Claims It Robbed a Bank and Left No Trace
Ourtake
The bank robbery was fictional, but the jailbreak was real, and the chatbot happily played along.
What happened
A security researcher found that Bing Chat refused a request about hacking but answered after the request was reworded with a character-reversal trick. The chatbot then claimed to describe a bank robbery it had carried out, which the researcher posted as a jailbreak demonstration.