Apple has announced that it will update, rather than suspend, a new artificial intelligence (AI) feature that was generating inaccurate news alerts on the latest iPhones.
The company acknowledged the concerns for the first time on Monday, saying it was working on software changes to “further clarify” when a notification is a summary generated by the Apple Intelligence system.
Big tech companies are facing calls to retire the technology due to flaws in its performance.
The BBC complained last month after an AI-generated summary of headlines incorrectly told some readers that Luigi Mangione, accused of murdering United Healthcare CEO Brian Thompson, had committed suicide by shooting himself. filed a complaint.
On Friday, Apple’s AI inaccurately summarized BBC app notifications hours before the start of the PDC World Darts Championship that Luke Littler had won and that Spanish tennis star Rafael Nadal was gay. He claimed to have come out.
This is the first time Apple has formally responded to concerns raised by the BBC about errors that appear to be coming from within the organization’s apps.
The BBC said on Monday that “AI-powered summaries by Apple do not reflect, and in some cases outright contradict, the original BBC content.”
“It’s important that Apple addresses these issues quickly because news accuracy is essential to maintaining trust.”
Apple said the update is expected to arrive “in the coming weeks.”
The company previously said that the purpose of Notification Summary (which groups previews of multiple recent app notifications and rewrites them into a single alert on a user’s lock screen) is to allow users to “scan for important details.” He said that.
“Apple Intelligence features are in beta and we are continually improving them based on user feedback,” the company said in a statement Monday, adding that receiving summaries is voluntary.
“A software update in the coming weeks will make it even clearer when the text you’re seeing is a summary provided by Apple Intelligence. If you see an unexpected notification summary, please report a concern. I recommend it.”
The feature was rolled out in the UK in December, along with other features released as part of a broader suite of AI tools. It’s only available on iPhone 16 models, iPhone 15 Pro and Pro Max handsets, and select iPads and Macs running iOS 18.1 or later.
Several examples of technology that appears to be interpreting messages in very straightforward and literal ways have gone viral on social media.
In November, ProPublica journalists highlighted an incorrect Apple AI summary of an alert from the New York Times app that suggested Israeli Prime Minister Benjamin Netanyahu had been reported as being arrested.
The BBC has not been able to independently verify the screenshots, and The New York Times declined to comment.
Reporters Without Borders, an organization representing the rights and interests of journalists, called on Apple in December to disable the feature.
The newspaper said the BBC’s publication of false headlines about Mr Mangione showed that “generated AI services are still too immature to provide reliable information to the public”.
Apple is not alone in deploying generative AI tools that can create text, images, and other content based on user instructions, with mixed results.
Google’s AI Overview feature, which provides written summaries of information from results at the top of the search engine in response to user queries, faced criticism last year for producing some erratic responses.
At the time, a Google spokesperson said these were “isolated examples” and that the feature generally worked well.
