In December 2024, the BBC filed a complaint with Apple for erroneous AI-generated headline summaries.
Apple has apparently deactivated its Apple Intelligence notification summaries feature for select apps in response to criticism over error-filled summaries for news items for beta software users, indicating ongoing issues with AI technology. The artificial intelligence (AI) program recently misrepresented a Washington Post news notification, drawing criticism from users and news outlets. Last month, the BBC stated that iOS 18 created an erroneous summary of a news piece. According to reports, the corporation has temporarily deactivated the feature in order to address difficulties before re-releasing it to the public.
The decision to stop AI summaries comes weeks after the BBC revealed that Apple’s AI system distorted its news notifications to reflect incorrect information. The break only affects users of Apple’s beta software, not those on the company’s major operating systems.
Last Thursday, news and entertainment apps like The New York Times began presenting a short notification inside the iPhone settings app informing users that AI-powered summaries were “temporarily unavailable.”
The hold on one of Apple Intelligence’s major functions demonstrates the difficulty Apple confronts in rolling out its artificial intelligence technology, which has been questioned by many social media users.
iOS 18.3 disables the app alert summary feature for news and entertainment apps and with the iOS 18.1 release where Apple added additional Apple Intelligence functions to compatible iPhone models. One of these was the notification summaries feature, which condenses the content in the message and allows users to rapidly glance over crucial elements. It is an opt-in feature that allows users to specify which apps’ alerts will be summarized.
Since its inception, the AI feature has been observed producing inaccurate summaries of notifications from news applications. In December 2024, the BBC contacted Apple regarding inaccurate information in AI summaries. This was purportedly done when Apple Intelligence falsely claimed that Luigi Mangione, the convicted murderer of the UnitedHealthcare CEO, committed suicide.
A Bluesky user also uploaded a screenshot in which the AI tool described a New York Times story and incorrectly claimed that Israel Prime Minister Benjamin Netanyahu had been arrested, despite the fact that the original news notice indicated that the International Criminal Court had issued a warrant for his arrest.
Despite these difficulties, Apple did not take any steps to remove the functionality or release a patch at the time. The most recent reported problem occurred on Wednesday, when the feature wrongly summarised a push notice from The Washington Post.
In response, Geoffrey Fowler, the publication’s technology columnist, wrote, “These mangled summaries were amusing on low-stakes text messages, but more dangerous on news app alerts.” One time, it mistakenly informed me that Donald Trump had endorsed Tim Walz for president.”
While presenting customers incorrect and error-laden summaries is a major concern, the problem is exacerbated by how Apple displays the AI summaries. Currently, aside from a little indicator, AI-generated notification summaries do not explicitly indicate when AI is used to summarize a notification.
According to 9to5Mac, iOS 18.3 notification summaries will display summarised material in italics. Meanwhile, AI summaries for news and entertainment applications will be disabled by default with the iOS 18.3 update.
As on iOS 18.3 beta 3, users will also be able to turn off notification summaries for an app from the lock screen or the notification center. Apple warns users that summaries “may contain errors” in the Settings app.
According to the magazine, Apple will strengthen the feature and offer notification summaries for the blocked app category in a subsequent software update.
“With the latest beta software releases of iOS 18.3, iPadOS 18.3, and macOS Sequoia 15.3, notification summaries for the News & Entertainment category will be temporarily unavailable,” an Apple spokeswoman said in a statement to CNBC.
According to the spokesman, Apple is working on software changes that will be released in a future update. The company did not specify when it would release iOS 18.3 to users of the main version of the iPhone operating system, but based on Apple’s previous software release patterns, it could take weeks.
The decision to temporarily suspend the AI summaries comes on the same day as Apple’s stock fell 4%, its least strong day since August 5. The dip was attributed to famous Apple supply chain expert Ming-Chi Kuo’s statement on Monday that the Apple Intelligence suite of capabilities does not appear to be increasing iPhone sales.
The difficulties of Apple Intelligence since inception was unveiled in October as the defining feature of its latest range of iPhone models and as a counter to Silicon Valley’s AI arms race, which began with the release of OpenAI’s ChatGPT in late 2022.
Apple has embraced AI technologies as a significant selling point in commercials and marketing for its latest hardware goods, but the software has been plagued by bugs.
The business claims that the entire Apple Intelligence system is in beta, and the update on Thursday included language stating that the AI software may deliver unexpected results.
Apple Intelligence has several features, including image generators, but the one that has received the most attention is its ability to summarize entire stacks of notifications into concise sentences — useful, according to Apple’s promotional items, for scrolling through hundreds of group chat notifications without scrolling through the entire conversation.
With the Thursday update, Apple said it will show any AI-summarized notice in italics to distinguish them from other notifications.
Apple Intelligence summaries were not flawless during testing, although the flaws were mostly humorous and obvious. Problems arose when the technology was employed to summarize news, and it revealed incorrect information.
The most egregious well-documented blunder occurred in December, when 22 distinct BBC news notifications were combined into a three-part headline beginning “Luigi Mangione shoots himself.” The supposed Brian Thompson assassin did not do so.
According to a ProPublica reporter’s social media post, the feature also blended New York Times headlines into a November message that erroneously claimed Israeli Prime Minister Benjamin Netanyahu had been jailed.
Another Apple Intelligence message on Jan. 3 stated that darts player Luke Littler had won a global championship that had yet to be held, according to the BBC. The system also combined messages from the BBC’s sports app to say that “Brazilian tennis player, Rafael Nadal, comes out as gay.” Nadal is Spanish and married to Maria Francisca Perello.
On Thursday, Apple also released a new feature that allows users to disable AI summaries for any app simply swiping left on the notification from the phone’s lock screen. Users could previously only disable AI summaries through the Settings app.
Discover more from TechBooky
Subscribe to get the latest posts sent to your email.