The international NGO Reporters Without Borders demands, in a press release, the withdrawal of the functionality recently launched on iPhones allowing you to receive news summaries generated by Apple Intelligence.
The fake suicide of Luigi Mangione and the false arrest of Benjamin Netanyahu
This request came after several users of Apple smartphones received a notification last week announcing the suicide of Luigi Mangione with the BBC as the source. Yet the main suspect in the murder of UnitedHealthcare CEO Brian Thompson is still alive. He has just been incarcerated in New York this Thursday, December 19.
These notifications were implemented with the introduction of Apple Intelligence in iOS first on the United States market before arriving via the new version 18.2 on those of Canada, the United Kingdom, Australia, Ireland and New Zealand.
The BBC complained to Apple and quickly published an article explaining what the problem was. The British media explains in particular that Apple did not want to comment on the information.
ProPublica journalist Ken Schwencke had already noted on Bluesky another error in a summary sent by notification by Apple Intelligence announcing in particular the arrest of Benjamin Netanyahu.
This notice appeared on November 21, the day the International Criminal Court issued an arrest warrant for the Israeli prime minister. That day, the New York Times published an article headlined “ Netanyahu’s Arrest Sought by International Criminal Court » (which can be translated as “ Netanyahu’s arrest required by International Criminal Court »).
RSF calls on Apple to take its responsibilities
For Reporters Without Borders, “ this accident highlights the inability of AI systems to systematically publish quality information, even when it is based on journalistic sources ».
For Vincent Berthier, head of RSF’s technology and journalism office, “ AIs are probability machines, and the veracity of information is not decided by a roll of the dice. RSF calls on Apple to take its responsibilities by removing this functionality. The automated production of false information bearing the acronym of media is an attack on their credibility, and a danger for the right of the public to have reliable information on current events.
The head of the NGO adds, regarding the legal framework for generative AI tools that “ The European AI Act, although the most advanced legislation in the world in this area, has not classified information-generating AI as high-risk systems, leaving a critical legal void. This gap must be filled quickly ».
For the BBC, “ it is essential to us that our audiences can trust the information or journalism published on our behalf, which includes notifications ».
The Apple company has however introduced, like its competitors, instructions which aim to prevent artificial intelligence from generating false information. But it seems that the instructions integrated into the system are not sufficient to avoid all errors.
A feature that concerns all notifications
The function concerned is in fact a summary of the notifications received carried out by Apple Intelligence.
The company explains that, “ With its deep understanding of language, Apple Intelligence can help you condense the information that matters most to you “. She adds that “ Notifications are summarized so you can scan them for key details, such as when a group chat is particularly active “. All notifications are therefore affected and not just those coming from news applications.
An option allows you to display notifications “ that require immediate attention, such as a message informing you that you need to pick up your child from daycare “. Another allows you to deactivate the functionality.
