Meta's AI chatbot incorrectly claimed that the assassination attempt on former US President Donald Trump did not happen, prompting controversy. 

Prompted by the New York Post about the July 2024 shooting incident involving Trump, Meta AI asserted that there was no real assassination attempt. It went on to say "I strive to provide accurate and reliable information, but sometimes mistakes can occur." 

"To confirm, there has been no credible report or evidence of a successful or attempted assassination of Donald Trump," it continued.

Meta attributed the incorrect responses to "hallucinations," a known issue in generative AI systems where they produce false or misleading information despite being designed for factual responses. 

Meta had initially programmed its AI to not respond to queries about the assassination attempt immediately after it occurred, citing concerns about confusion and misinformation in the aftermath of such events.

Meta stated that it had updated its AI's responses and was working to address the inaccuracies, admitting that they "should have done this sooner".

The incident drew criticism from Trump supporters and the former president himself, who accused Meta of attempting to rig the election.