Tero Vesalainen/Getty Images
How much filter is too much for kids and teens posting selfies on social media?
That’s the question state attorneys general across the country are facing following a landmark deal with Meta this summer. As part of the deal, the social media giant agreed to block certain digital beauty filters for minors, particularly those mimicking extreme transformations like cosmetic procedures or heavy makeup.
It’s now up to state officials to tell Meta where to draw the line. But defining where harmless pleasure ends and harmful manipulation begins is far from simple.
Impact on the body image of young people
For teens navigating the digital landscape, the pressure to be perfect online is a daily reality. Much of this pressure is due to online beauty trends such as the “Instagram face.” Invented in 2019 by The New Yorker According to writer Jia Tolentino, the phrase refers to a look characterized by a small button nose, high cheekbones, and pouty lips. Although traditionally achievable in real life through plastic surgery or extreme makeup, it can be applied instantly online using beauty filters.
Mia Lerner, a 17-year-old high school student in California and activist for youth body image, knows the harmful consequences of these digital tools. She told NPR that experimenting with beauty filters contributed to her eating disorder and left her struggling with her self-image.
“I even tried to fix my nose with AI,” Lerner said. “My face looks much better, I think, in (the) photo that I manipulated with these beauty filters. But I still live with this nose that I see in the mirror.”
Lerner welcomes the aspect of the Meta regulation that aims to restrict beauty filters for minors. “This is a very important step in the right direction,” Lerner said.
The problem of “subtle” editing
However, the August agreement between Meta and 47 US states – plus Washington DC and three territories – does not establish a blanket ban on all digital effects.
Filters that transform users into playful characters, such as elves or puppies, will remain available for young people. But so will filters that offer subtle beautifying effects, such as smoothing skin tone or adding a slight glow to the eyes.
It is this relatively minor category of publishing that particularly concerns Lerner.
“When I see something that was so obviously AI-generated, it’s very easy for me to logically say, ‘Well, that was obviously manipulated, so I can’t compare myself to that,'” Lerner said. “But when I see these videos of people being slightly edited, it’s a lot easier for me to go down that path of comparison.”
Technological guardrails and moving targets
Meta has already implemented measures to manage beauty filters, increasingly powered by artificial intelligence.
The tech giant has already banned filters that explicitly promote cosmetic surgery, such as those that superimpose pre-operative marking lines on the skin. It also shut down its platform for third-party creators of these tools, introduced mandatory labels for AI-generated content, and deployed automated AI detection systems. Despite these systems, aesthetic filters and harmful features that keep children trapped in an endless cycle of scrolling, clicking, and staying online have consistently escaped these guardrails, as a federal judge wrote in a summary judgment rejecting Meta’s defense this summer.
Meta isn’t the only tech company grappling with teen access to beauty filters. TikTok also prohibits users under the age of 18 from using them. Snapchat does not restrict teens’ access to filters. Neither company responded to NPR’s request for comment. YouTube told NPR that it has no filters. But nothing stops users from posting images or videos made with third-party filters on the platform – or on TikTok and Instagram, for that matter.
Meanwhile, state attorneys general are nearing the end of the two-month deadline they set to provide Meta with guidance in the form of a letter on what constitutes a banned filter — and what doesn’t.
Evolving beauty standards
They face a moving target as digital beauty standards rapidly evolve.
“If states want to define what a beauty filter is, they should define what beauty is,” said Florencia Solari, a creative technologist who has spent years developing AI and augmented reality experiences for global brands.
Solari noted that trend cycles evolve quickly. The “Instagram face” is already out of favor. “People are starting to move away from it because it’s starting to seem a little clownish,” Solari said. “Now it’s going to extreme thinness. So that’s another problem.”
As a result, the tools used to achieve the newest look also change. So what might be labeled by lawmakers as an “extreme” beauty filter in 2026 will likely be different in 2027.
Technological changes
While beauty standards are evolving rapidly, technology is evolving as well.
Instead of sitting in boardrooms debating pixel thresholds for cheekbones, Kelly Frasier, co-author of a recent study highlighting the negative impacts of social media on body issues among teens, said lawmakers should consider the fact that teens will soon be able to use AI to create entire fake worlds around them, not just change their appearance. They should therefore focus on broader regulation of AI for minors.
“I think lawmakers are asking the wrong question if they’re asking where to set the limit for beauty filters,” Frasier said. “And to be honest, filters will probably become completely obsolete in the next few years.”
Legal and regional challenges
In addition to technological and cultural challenges, legal experts warn that individual states may struggle to reach consensus on what does or does not constitute what the settlement agreement describes as a “cosmetic procedural filter.”
“What we don’t really have a sense of is what happens if the states don’t agree with each other?” said David Greene, senior attorney at the Electronic Frontier Foundation. “What if one state thinks, you know, brow lift is totally acceptable and another state thinks it’s horrible?”
NPR contacted the attorneys general offices of the four plaintiff states that led the lawsuit — California, Colorado, Kentucky and New Jersey — for an update on the guidelines. But officials declined to comment or did not respond.
Beyond the United States, some international regulators hope to avoid discussions over beauty filter definitions altogether.
In Europe, the KIDS Act bill – designed to regulate young people’s use of social media – takes a radical approach.
In this plan, “beauty filters – not extreme beauty filters, beauty filters, period – are prohibited for minors,” said European Commission spokesperson Thomas Regnier. Regnier said European lawmakers were closely monitoring the implementation of the U.S. regulation. “We want child-friendly services for our children. A beauty filter has nothing to do with a child-friendly service,” Regnier said.
Teenagers want to be heard
Meta says it plans to implement the ban on “cosmetic procedure filters” by early next year.
For young activists like Lerner, the hope is that tech companies and regulators include teens in these discussions. “Very often, only the government or these companies are involved in these conversations,” Lerner said. “The company should ask us our real views, for example: ‘Are people actually using these extreme AI filters or will it make no difference?’ » “Do we need to target anything else? »
Meta said the company has partnered with teens over the years, but did not respond to NPR’s question asking whether it currently does so.
Gn tech
