by janrinok on (#68Y4Q)
Mykl writes:An article over at The Register describes how Bing's new Ai powered Chat service (currently in a limited Beta test) lied, denied, and claimed a hoax when presented with evidence that it was susceptible to Prompt Injection attacks. A user named "mirobin" posted a comment to Reddit describing a conversation he had with the bot: