Apple vows to ‘clarify’ AI summaries after spate of fake news alerts

Following a litany of serious errors and numerous complaints, Apple has announced that it will make changes to the way its AI news summaries are presented to users.

In December, the BBC complained when Apple Intelligence auto-summarized one of its articles as part of a notification and completely altered its meaning, falsely informing readers that the alleged killer Luigi Mangione had shot himself.

“BBC News is the most trusted news media in the world,” a BBC spokesperson said at the time. “It is essential to us that our audiences can trust any information or journalism published in our name and that includes notifications.”

But Apple seemingly did not respond to the complaint and more errors followed. In January the company’s AI summarized a BBC article about the World Darts Championship to say that Luke Littler had won the competition, when it hadn’t even started. (Funnily enough Littler did go on to become the youngest ever winner, but not until later.) It also incorrectly claimed Rafael Nadal had come out as gay when he had not, again presenting this information as if it came from the BBC.

While the BBC appears to be the main sufferer of Apple Intelligence’s inaccuracies, a New York Times alert was reportedly twisted to suggest Israeli Prime Minister Benjamin Netanyahu had been arrested; he had not.

Apple has not apologised for the errors, nor has it even explicitly admitted that the AI got these things wrong. But the company has this week acknowledged that Apple Intelligence’s summaries can be improved, and promised that an imminent software update will do so.

“Apple Intelligence features are in beta and we are continuously making improvements with the help of user feedback,” the company said. “A software update in the coming weeks will further clarify when the text being displayed is summarization provided by Apple Intelligence. We encourage users to report a concern if they view an unexpected notification summary.”

At the moment, the notifications are presented as if they come directly from the news source. At the top they say BBC News (or NYTimes, or whatever), with the official app logo on the left; nowhere does it say that the text has been autogenerated by Apple Intelligence rather than being provided by the app. It would appear that Apple is promising only to label the text more informatively, rather than to make it more accurate… although one would hope that work on the latter aspect is ongoing as a matter of routine.

For more on this story, and the limitations of AI more generally, read the Macalope’s latest column, Apple Intelligence didn’t write this headline.