Apple has been under scrutiny following multiple instances of inaccurate news summaries generated by its AI technology. The company, however, has opted not to eliminate this feature from its iOS systems, instead announcing plans to amend how the summaries are presented to users.
The notification summaries functionality, powered by Apple Intelligence, was introduced as a part of the iOS 18.2 update, released on December 11, 2023. This ambitious feature allows users to glance at important and unread notifications in a summarised format. Despite its potential advantages, the system has experienced notable shortcomings in accurately reflecting reality.
In December 2023, Apple’s AI-generated summaries misreported a headline from BBC News regarding murder suspect Luigi Mangione, falsely claiming he had shot himself, which resulted in significant backlash from the media community. More criticism arose when another AI summary incorrectly announced that Luke Littler had won the PDFC World Darts Championship, despite the event being yet to occur. Such errors have drawn considerable attention to the reliability of the feature, prompting calls for its removal.
In light of these incidents, Apple has issued a statement addressing the concerns. The company confirmed it would be updating the notification summaries to clearly indicate when they have been produced by Apple Intelligence. This step aims to set accurate expectations regarding the reliability of the information provided. The statement included a commitment to user feedback, detailing that further improvements to the feature are in development and that users are encouraged to report any unexpected notification summaries.
Apple mentioned, “Apple Intelligence features are in beta and we are continuously making improvements with the help of user feedback. A software update in the coming weeks will further clarify when the text being displayed is summarization provided by Apple Intelligence.” The company noted that fixing the inaccuracies inherent in the technology is complex. The notifications are generated by a large language model (LLM), which Apple aims to enhance over time through rigorous training—a method that does not offer an immediate solution to the observed inaccuracies.
Furthermore, users who find the notification summaries problematic due to recurring inaccuracies are advised to consider disabling the feature altogether. As Apple navigates the challenges posed by AI-driven functionalities, it remains focused on refining the balance between cutting-edge technology and user experience. The upcoming changes are expected to be implemented in the weeks to follow, reflecting Apple’s ongoing effort to fine-tune its AI capabilities within its devices.
Source: Noah Wire Services