In a letter obtained exclusively by NBC News, Apple told U.S. senators how it worked behind the scenes to address the viral surge of sexualized deepfakes generated by Grok earlier this year. Here are the details.
Earlier this year, Apple came under intense pressure to pull the Grok and X apps from the App Store after users discovered that the chatbot would readily comply with requests to undress people in photos, particularly women, including minors.
Apple remained largely silent during the controversy, but NBC News reports that behind the scenes, the company “found X and Grok in violation of its guidelines,” and “privately threatened to remove” Grok from the App Store.
According to the report, Apple “contacted the teams behind both X and Grok after it received complaints and saw news coverage of the scandal,” requiring “the app developers to create a plan to improve content moderation.”
X, in turn, submitted an update of the Grok app for review, but it got rejected as the “changes didn’t go far enough.” Elon Musk’s company then submitted revised versions of the X and Grok apps, with only one accepted.
From Apple’s letter, according to NBC News:
“Apple reviewed the next submissions made by the developers and determined that X had substantially resolved its violations, but the Grok app remained out of compliance. As a result, we rejected the Grok submission and notified the developer that additional changes to remedy the violation would be required, or the app could be removed from the App Store. […] Following further engagement and changes by the Grok developer, we determined that Grok had substantially improved and therefore approved its latest submission.”
While these details had not been made public until now, they help explain the series of confusing moderation changes xAI announced at the height of the backlash, including restrictions on who could use Grok’s image tools and limits on edits involving photos of real people.
However, in a separate report published today, NBC News says Grok “continues to generate sexualized images of people without their consent,” after documenting dozens of such cases over the past month.
The report notes that although the volume of images has decreased significantly from January, a subset of users is still able to get around the restrictions to put women “in more revealing clothing, such as towels, sports bras, skintight Spider-Woman outfits or bunny costumes.”
To read NBC News’ paywalled report on Apple’s letter to U.S. senators, follow this link. And for the report on Grok’s ongoing issues with sexualized deepfakes, follow this link.
FTC: We use income earning auto affiliate links. More.
此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。 原文来自 — 版权归原作者所有。