Meta accused of bias after AI chatbot fails to acknowledge Trump assassination attempt

Joel Kaplan, Meta’s global head of policy, described the AI-generated responses about the shooting as “unfortunate” and confirmed that “in a small number of cases, Meta AI continued to provide incorrect answers, including sometimes asserting that the event didn’t happen.”

He also said that the company was quickly working to address the issue. Kaplan said that Meta had initially recalibrated the AI chatbot to avoid answering questions about the Trump shooting. However, these restrictions were later removed after users started reporting the issue.

“These types of responses are referred to as hallucinations, which is an industry-wide issue we see across all generative AI systems, and is an ongoing challenge for how AI handles real-time events going forward,” Kaplan said in a blog post published on Tuesday, July 30.

Kaplan also revealed that a real photo of Trump at the site of the shooting was incorrectly tagged with a fact-check label. “Given the similarities between the doctored photo and the original image – which are only subtly (although importantly) different – our systems incorrectly applied that fact check to the real photo, too. Our teams worked to quickly correct this mistake,” he wrote in the blog.

Meanwhile, Google has also been accused of censoring search results about the failed assassination attempt via its autocomplete feature, according to a report by The Verge.

As US voters head to polling booths in a few months, there are apprehensions that the widespread availability of generative AI tools could fuel the spread of misinformation and lead to greater political manipulation.

The ability of social media platforms to combat AI-generated fake news has been called into question too. For instance, X’s policies against synthetic and manipulated media were violated by Elon Musk, the owner of the platform, when he shared a deepfake video of US Vice President and likely Democratic Party candidate Kamala Harris.