Reporters Without Borders (RSF) has urged Apple to remove its new generative AI feature after the BBC had asserted it created a misleading headline about a high-profile murder case in the United States.
The BBC announced on Thursday that it has filed a complaint against Apple after its AI tool, Apple Intelligence, which summarizes and groups notifications using artificial intelligence, generated a false headline about Luigi Mangione who has been charged with the stalking and murder in New York City of UnitedHealthcare CEO Brian Thompson. The BBC asserted that Apple’s summary incorrectly implied that the news company had published an article claiming that Mangione had shot himself—a claim it said is untrue.
Citing this complaint, RSF said it is seriously concerned about the risks AI tools pose to the credibility of media outlets, stating that “the BBC incident demonstrates that generative AI technology is still too unreliable to provide accurate information.”
Vincent Berthier, head of RSF’s technology and journalism desk, criticized the reliability of AI-generated summaries. “AIs are probability machines, and facts cannot be determined by chance,” he said.
Apple has not commented on the controversy since the story emerged last week.
The misleading notification appeared in a grouped summary that also included accurate headlines about the potential overthrow of Bashar al-Assad’s regime in Syria and a South Korean police raid involving President Yoon Suk Yeol.
The BBC stated that it had contacted Apple to address the issue but did not indicate whether the company had responded.
The BBC misquote is not the first instance of misleading AI-generated summaries. On November 21, the New York Times experienced a similar issue when Apple Intelligence grouped three unrelated articles into one notification. The notification falsely stated, “Netanyahu arrested,” in reference to Israeli Prime Minister Benjamin Netanyahu. The article was about an arrest warrant issued by the International Criminal Court, not Netanyahu’s arrest.
A journalist flagged the error and shared a screenshot on Bluesky. The journalist later confirmed the image’s authenticity, but the New York Times has declined to comment on the matter.
Apple Intelligence, introduced in the UK last week, allows users to group notifications to minimize interruptions. The feature is available on devices running iOS 18.1 or later, including all iPhone 16 models, the iPhone 15 Pro, and 15 Pro Max, as well as certain iPads and Macs.
Grouped notifications are marked with a distinctive icon, and users can report issues directly through the interface. Despite its convenience, the feature has faced criticism for its inaccuracies, raising questions about the readiness of generative AI for public use.