Disable auto-play and infinite scroll or risk massive fines, EU tells Meta

If Meta fails to make changes to comply with the EU’s Digital Services Act, the company risks fines up to 6 percent of its global annual turnover when the EC makes its final decision in the coming months. “Our starting point is that, based on our findings, this design is too addictive and changes need to be made,” Henna Virkkunen, the EU’s tech chief, told Reuters. “The next step is either that Meta changes its design or a non-compliance decision will follow,” she said, noting in the press release that the EU’s priority is “protecting the physical and mental health of Europeans.” Virkkunen said the EU does not plan to back off the fight if the final decision reiterates the preliminary findings. “The Digital Services Act provides a clear framework to hold platforms accountable for the addictive design and effects of their services,” Virkkunen said. “We are fully committed to enforcing our legislation in Europe.” As Meta seemingly continues treading water—pointing to screen-time notifications that kids can easily dismiss or default settings that may be changed—the financial pressure to do more to protect kids could threaten its AI ambitions at a crucial time. Right now, Meta is spending a fortune, The Information reported, rushing to advance its AI to catch up with rivals. At a recent internal town hall, Meta claimed its upcoming AI model, codenamed Watermelon, had “caught up with OpenAI’s flagship GPT-5.5 model,” sources familiar with the meeting’s agenda told Business Insider. Meta’s uptick in investments came after struggling to “convince developers and customers that its models belong at the industry’s leading edge,” Business Insider reported. According to The Information, Mark Zuckerberg has already allotted “between $125 billion and $145 billion on capital expenditures for its AI data centers this year, in addition to significantly higher operating expenses, also thanks to AI, on outside cloud services and AI talent.” “How much money exactly is Zuckerberg willing to fork over on this quest?” The Information posited, while noting that since 2020, Meta’s Reality Lab division has lost $87 billion chasing metaverse pipe dreams. Meta’s AI strategy appears to depend on getting as much buy-in as possible for its AI tools. Reports indicate that its plan is to offer models at “dirt cheap” prices to “undercut rivals,” then possibly increase prices once adoption peaks. But Meta’s scramble to make its AI offerings more appealing is clashing with concerns that users already have about its platforms, which allegedly addict users to generate endless data that now fuels its AI models. This week, Meta faced backlash after releasing a new AI model, Muse, which mines public Instagram feeds for images and videos. NBC News tests found that Muse can be used to create deepfakes of celebrities and regular people, which Meta’s deepfake detection tools don’t always catch. That heightens the risk that Facebook and Instagram users will be targeted in deceptive deepfakes, NBC News reported. When announcing the new model, Meta confirmed that the majority of platform users were automatically opted in, with the platform seemingly hoping to seize as much data as possible from the start. The only exceptions were made for users whose profiles were already set to private and users under 18 whose sharing and reuse settings were toggled off by default. Any content sucked into Muse will be discoverable on search engines, Meta confirmed, while positioning the AI tool as a new way to create content and increase engagement on its platforms. If Meta platforms were left as is, it’s easy to see how its personalized recommendations could drive views of this content, over time increasing its use and expanding Meta’s AI capabilities. Meta has not clarified why it had to opt users in to sharing by default but said in a statement to Reuters that it’s easy to opt out in a “few clicks.” For example, by navigating to “sharing and reuse” Instagram settings, users can toggle off options to “allow people to create with and reuse your content” and “allow people to create with and reuse your original audio on Meta AI.” Many users rushed to share instructions on how to opt out of the sharing, as privacy advocates raised concerns that Meta did not seek direct consent. Similarly child safety advocates noted that kids can still opt in to allow people they follow to use their content in ways that could possibly feed sensitive data into Meta’s models or make sensitive images discoverable. There may also be minors on the platform evading age verification whose data could be used.