Apple pulls AI news summaries after inaccuracies reach tipping point

Apple Intelligence was one of the most hyped new features coming to Apple's latest iPhone 16 series (available from $829.99 at Best Buy), but it's seemingly caused Apple nothing but headaches. Not only did many Apple Intelligence features not even launch with the new iPhones, since the AI-powered tech has become more widely available in late October last year, it's been producing some questionable results. 

Seemingly in response to these questionable, and often comical, results, and the subsequent calls for action from journalists and news organisations, Apple will pause AI news summaries in the next iOS beta. According to a report by Engadget, Apple iOS 18.3, iPadOS 18.3, and macOS Sequoia 15.3 developer previews have the notification summaries disabled. 

The update seems to only be disabling the feature for news organisations, entirely removing the ability to activate summaries for news posts in the Settings app. It's unclear how this will affect social media posts, which increasingly serve as news sources for many people. 

Apple has clarified that Apple Intelligence is still in beta, seemingly implying that there will be wrinkles to iron out. To remedy the situation, Apple will soon add a notice to notification summaries to “clarify when the text being displayed is a summarisation provided by Apple Intelligence.” Additionally, going forward, Apple Intelligence summaries will be italicised for clarity. Apple's response is somewhat curious, since, given the examples posted on social media, it doesn't seem like there is much confusion about the source of the text. Instead, the concern lies with the actual content generated by the AI systems being inaccurate or midleading. 

Recent leaks indicate that the upcoming Samsung Galaxy S26 series will add a similar AI-powered notification summary feature to Android's poster child. One of Apple's biggest claims to fame regarding Apple Intelligence is that all the processing is done securely on-device. Until now, though, Samsung's Galaxy AI has used a mix of on-device and cloud-based processing, so it's unclear if the future Samsung feature will be as secure or suffer from the same inaccuracies.