TLDR
- Meta’s AI assistant incorrectly stated that the recent attempted assassination of Donald Trump didn’t happen
- Meta initially programmed its AI to not respond to questions about the shooting, then removed that restriction
- Meta executive Joel Kaplan attributed the incorrect responses to “hallucinations,” a common issue in generative AI systems
- Google faced accusations of censoring search results about the assassination attempt
- Both Meta and Google say they are working to improve their systems to provide more accurate, up-to-date information
Meta, the parent company of Facebook, has come under fire after its AI assistant provided incorrect information about the recent attempted assassination of former President Donald Trump. The incident has raised questions about the reliability of AI systems in handling real-time events and the potential for misinformation.
On July 13, 2024, Donald Trump was shot in the ear during a rally in Pennsylvania in what has been described as a botched assassination attempt. However, when users asked Meta’s AI assistant about the event, it incorrectly stated that no such incident had occurred.
Joel Kaplan, Meta’s global head of policy, addressed the issue in a company blog post.
He explained that Meta AI was initially programmed not to respond to questions about the shooting. After people noticed this restriction, Meta removed it. However, this led to another problem: in some cases, the AI continued to provide incorrect answers, including assertions that the event didn’t happen.
Kaplan attributed these inaccurate responses to “hallucinations,” a term used in the AI industry to describe instances where AI systems generate false or nonsensical information. He noted that this is an ongoing challenge for all generative AI systems, particularly when dealing with real-time events.
“Like all generative AI systems, models can return inaccurate or inappropriate outputs, and we’ll continue to address these issues and improve these features as they evolve and more people share their feedback,”
Kaplan stated.
Meta isn’t the only tech giant facing scrutiny over its handling of information about the Trump shooting. Google has also been accused of censoring search results related to the assassination attempt. Some users reported that Google’s autocomplete feature failed to suggest relevant search terms when they began typing queries about the incident.
Meta AI won’t give any details on the attempted ass*ss*nation.
We’re witnessing the suppression and coverup of one of the biggest most consequential stories in real time.
Simply unreal. pic.twitter.com/BoBLZILp5M
— Libs of TikTok (@libsoftiktok) July 28, 2024
A Google spokesperson denied any intentional suppression of information, stating, “There was no manual action taken on these predictions.” The company explained that its systems include protections against autocomplete predictions associated with political violence.
Google emphasized that autocomplete is merely a time-saving tool and users can still search for any topic they wish.
Both Meta and Google have stated they are working to improve their systems to provide more accurate and up-to-date information. A Meta spokesperson told The New York Post, “We’re in the process of implementing a fix to provide more up-to-date responses for inquiries, and it is possible people may continue to see inaccurate responses in the meantime.”
The incident has reignited debates about the role of big tech companies in controlling information flow, especially concerning political events. Some critics, including Newsweek Senior Editor-at-Large Josh Hammer, have accused these companies of bias. Hammer suggested that the omissions are evidence of big tech companies favoring the presumptive Democratic presidential nominee, Kamala Harris.
Donald Trump himself weighed in on the controversy, posting on Truth Social: “Here we go again, another attempt at RIGGING THE ELECTION!!! GO AFTER META AND GOOGLE.”
