Apple has pulled a new iPhone feature released just three months ago after users slammed it for spreading misinformation.
The tech giant removed its AI notification summaries for news and entertainment apps after the system falsely reported a news article.
The summary of the BBC article suggested that Luigi Mangione, 26, the alleged assassin of the CEO of UnitedHealthcare, had shot himself.
It read: ‘Luigi Mangione shoots himself; Syrian mother hopes Assad pays the price; South Korea police raid Yoon Suk Yeol’s office,’ in reference to three articles that had supposedly been published by the BBC.
Mangione has been accused of shooting Brian Thompson, 50, at point-blank range as he was walking to a Manhattan hotel where his company was holding an investor conference on December 4. He is currently being held in a Brooklyn federal jail.
Apple said it is disabling the AI-generated feature for news and entertainment while it tries to fix the issue causing the technology to fabricate information — a problem often described as ‘hallucinations’ within the industry.
Entertainment and news apps are now displaying a message that states AI-powered summaries are ‘temporarily unavailable,’ inside of the iPhone settings app.
The feature was canceled in the iOS 18.3 beta software but is said to roll the update out to everyone by the end of the month. Even if it is just temporary, the suspension represents a blow to Apple’s efforts to bring AI to the iPhone and its other products.
Apple Intelligence launched on October 28, 2024, unleashing new technologies to the iPhone 15 Pro models and iPhone 16 family.
The tech giant announced Apple Intelligence with great reverence as ‘the personal intelligence system that combines the power of generative models with personal context to deliver intelligence’.
To put it much more simply, it’s an umbrella term for all the jobs AI can do on your iPhone, iPad and Mac computer – as long as it has the upcoming iOS 18 software and an all-important AI chip inside the device.
Tucked in the tech are AI writing tools that write, proofread and summarize text on multiple apps, including news apps.
The decision to axe the feature was disclosed Thursday as part of a test version for the next software release, iOS 18.3.
The beta version is only available to a relatively small group of iPhone users and developers, but the same features are usually released in an update available to all users several weeks after the testing begins.
After identifying the error, a spokesperson for the BBC contacted Apple ‘to raise this concern and fix the problem,’ according to the broadcaster.
The BBC pointed out that this isn’t the first time Apple Intelligence has given misleading summaries of news articles.
On November 21, the New York Times suffered from a similar mistake.
The AI grouped three articles together in one notification, with the first part saying ‘Netanyahu arrested,’ referring to the Israeli prime minister.
The prime minister has not been arrested, nor did the newspaper incorrectly report that he has been.
The mistakes appear to be relatively widespread.
Numerous iPhone users have shared screenshots of notification summaries that are inaccurate and sometimes nonsensical.
‘Big fan of Apple Intelligence’s summary feature—mainly because it turns every boring notification I wouldn’t previously read into a cute little mystery to unwrap,’ one X user posted with a screenshot of an Apple News summary.
The summarized notification read: ‘Love salmon might not be a good idea; polar bears are back in Britain.’
This feature doesn’t just summarize news articles. It can also summarize your messages, sometimes with alarming inaccuracy.
Another X user posted a screenshot of Apple Intelligence’s attempt to summarize a text from their mom.
It read: ‘Attempted suicide, but recovered and hiked in Redlands and Palm Springs,’ when the actual message apparently read: ‘That hike almost killed me!’
These screenshots are unverified, but they are among many other reports of similar issues with the Apple Intelligence feature. While some of them may be easy to laugh at, there are dangers associated with misreporting facts in this way.
Prof Petros Iosifidis, a professor in media policy at City University in London, told BBC News that there are potential advantages to this type of feature, ‘but the technology is not there yet and there is a real danger of spreading disinformation.’
‘I can see the pressure getting to the market first, but I am surprised that Apple put their name on such demonstrably half-baked product,’ he said.
An Apple spokesperson noted that Apple is working on fixing issues with the software that will be seen in a prospective update, but did not specify when the update will be.
