Journalism group urges Apple to disable AI summaries after fake headline incident

What just happened? Just days after Apple's AI-powered notification summary tool pushed out a false BBC headline about Luigi Mangione, a major trade body is urging the company to remove the feature completely. It marks the latest setback in Apple's attempts to convince its customers that its AI is worth using.

On December 13, Apple Intelligence, which has a history of making significant errors when it comes to summarizing notifications, pushed out a summary of several BBC headlines that included the claim Mangione had shot himself. The incident happened just 48 hours after Apple Intelligence launched in the UK.

Courtesy of BBC News

Mangione, who has been arrested and charged with the first-degree murder of UnitedHealthcare CEO Brian Thompson in New York, has not shot himself. He remains in custody at Huntingdon State Correctional Institution in Huntingdon County, Pennsylvania.

It certainly isn't the first time Apple Intelligence has gotten a summary notification wrong. It previously claimed that Israeli prime minister Benjamin Netanyahu had been arrested after the International Criminal Court issued an arrest warrant.

Reporters Without Borders (RSF), an international non-profit NGO that focuses on safeguarding the right to freedom of information, is calling on Apple to disable the notification summary feature.

The RSF writes that the incident illustrates how generative AI services are still too immature to produce reliable information for the public, and should not be allowed on the market for such uses.

The group added that the probabilistic way in which AI systems operate automatically disqualifies them as a reliable technology for news media.

"AIs are probability machines, and facts can't be decided by a roll of the dice. RSF calls on Apple to act responsibly by removing this feature. The automated production of false information attributed to a media outlet is a blow to the outlet's credibility and a danger to the public's right to reliable information on current affairs," said Vincent Berthier, Head of RSF's Technology and Journalism Desk.

"The European AI Act – despite being the most advanced legislation in the world in this area – did not classify information-generating AIs as high-risk systems, leaving a critical legal vacuum. This gap must be filled immediately."

The BBC contacted Apple when it learned of the false headline to raise concerns and ask the company to fix the problem.

The summary notification in question showed three headlines: the fake one about Mangione, and two correct headlines, about the overthrow of Bashar al-Assad's regime in Syria and an update on South Korean President Yoon Suk Yeol.

Apple says its AI's summarization ability allows users to scan long or stacked notifications with key details right on the Lock Screen, such as when a group chat is particularly active. It often gets the summaries wrong or fails to understand their context, sometimes in hilarious fashion.

Spewing out false news headlines is always going to get Apple in trouble, especially at a time when companies are trying to convince people that AI is the future of pretty much everything. Apple has yet to respond to the incident, but don't expect it to permanently remove the feature. At most, Cupertino might disable it for a while.