State attorneys general are working out which digital beauty filters Meta must keep away from teenagers under the company’s landmark settlement with the states, and the line is proving hard to draw, NPR reported Oct. 25.

Meta agreed in August to pay up to $17.1 billion over 10 years and to add child-safety features to Facebook and Instagram to settle claims by 47 states, the District of Columbia and three territories that its platforms harm young users. Pennsylvania is expected to receive at least $516 million and as much as $729 million, according to Attorney General Dave Sunday. Only Florida, New Mexico and Texas are not part of the agreement, the Pennsylvania Capital-Star reported.

One provision bars minors from using what the settlement calls “cosmetic procedure filters.” The agreement defines those as digital filters or augmented-reality effects that distort, sculpt, redefine or idealize a user’s face in ways that could not be achieved without cosmetic surgery or extreme makeup, according to the Electronic Frontier Foundation’s review of the document. The states agreed to send Meta a letter with examples of effects that do and do not qualify within two months of the settlement taking effect. NPR reported that the states are nearing the end of that window, and Meta has said it plans to enforce the ban by early next year.

The settlement does not ban every effect. Playful filters that turn users into characters such as elves or puppies will stay available to young people, NPR reported, as will subtle beautification effects such as smoothing skin or adding a slight glow to the eyes.

That subtler category worries Mia Lerner, a 17-year-old high school senior in California who campaigns on youth body image. Lerner told NPR that experimenting with beauty filters contributed to her eating disorder and her struggles with self-image. She called the settlement provision “a very important step in the right direction,” but said lightly edited videos make it easier to compare herself to others than images that are obviously manipulated.

Meta already bans filters that explicitly promote cosmetic surgery, such as effects that draw pre-surgery marker lines on skin, and it has shut down its platform for outside creators of augmented-reality effects and added labels for AI-generated images, NPR reported. A federal judge nonetheless wrote this summer that harmful filters and features designed to keep children scrolling had slipped past those safeguards. Court filings in a Massachusetts case previously alleged that Chief Executive Mark Zuckerberg vetoed a 2019 proposal to disable Instagram’s beauty filters.

Other platforms handle the issue differently. TikTok restricts users under 18 from beauty filters, while Snapchat does not, according to NPR. YouTube told NPR it offers no filters, though users can upload videos made with outside tools to any of the platforms.

Experts quoted by NPR said the states face a moving target. Florencia Solari, a creative technologist who builds AI and augmented-reality experiences, said that if states want to define a beauty filter, “they should define what is beauty,” and noted that trends are already shifting away from the so-called “Instagram face.” Kelly Frasier, co-author of a study on social media and adolescent body image, argued that lawmakers should focus on broader AI rules for minors, predicting that filters will probably become obsolete within a few years. David Greene, senior counsel at the Electronic Frontier Foundation, questioned what would happen if states disagree about which effects cross the line.

European regulators are taking a broader approach. A proposed European Union law known as the KIDS Act would forbid all beauty filters for minors, not only extreme ones, European Commission spokesperson Thomas Regnier told NPR, adding that EU officials are watching how the U.S. settlement is carried out.

The four states that led the trial, California, Colorado, Kentucky and New Jersey, declined to comment or did not respond when NPR asked about the guidance. Their letter to Meta defining which filters must be blocked is the next step before the company begins enforcing the ban.

Sources