apple The company has reportedly disabled artificial intelligence-generated news summaries in the beta version of its new iOS operating system.
The company took this action after a previous update failed to prevent the issue. apple intelligence– The Washington Post is improving its ability to prevent information from being incorrectly summarized or simply fabricated. reported Friday (January 17th).
Some news organizations have called on Apple to do something about the feature after failing to claim summaries of news articles as their own.according to the report.
The latest update, iOS 18.3, is coming soon to all iPhones that support Apple Intelligence.says the report..
Apple aims to improve News feature and bring it back in a future updateaccording to the report.
reported In November, Apple announced that it was facing growing criticism over inaccurate and misleading headline summaries it had produced. News summary using AI Features.
The BBC was alerted to the issue in December when a breaking news summary, believed to be from the BBC app, incorrectly reported that Luigi Mangione, who is accused of murdering United Healthcare CEO, had shot himself. uttered.
In another example, Apple’s news summary stated before the contest that Luke Littler had won the PDC World Darts Championship. flat It has begun.
Other publications have seen similar hallucinations from Apple’s AI.
Vincent BerthierHead of Technology and Journalism Desk Reporters Without Borders (RSF in French) and called on Apple to remove the feature from iPhones.
“The automated production of false information by news organizations is a blow to their credibility and a danger to the public’s right to reliable information about current events,” Berthier wrote on the group’s website. said.
A.I. hallucinationPYMNTS reported in June that AI systems generate plausible but inaccurate information, creating troubling problems for companies.
As companies increasingly rely on AI to drive decision-making, the risks posed by these fabricated outputs have come into focus.
In October, OpenAI‘s whisper Transcription software caught add fabricated text Conversations containing potentially harmful content that the speaker never said.
On Tuesday (January 14th), reported the problem hallucination is one of the obstacles Amazon Encountered during the renovation planning of alexa to smarter A voice assistant with generative AI at its core.
For all of our PYMNTS AI coverage, subscribe to our daily AI newsletter.
