The demand for AR beauty filters on Instagram remains high, despite evidence of harmful effects on users' mental health and Meta's efforts to ban some filters
In October 2021, Facebook announced a massive pivot, changing its name to Meta and going all in on augmented and virtual reality through … Tweets: @techreview , @mattnavarra , @aarieff , and @hypervisible Tweets: @techreview : Demand for face filters on Instagram and other visual platforms remains high, despite their negative effect on users' mental health and Meta's efforts to ban some of them. https://www.technologyreview.com/ ... Matt Navarra / @mattnavarra : “The demand for deformation beauty filters is so consistent that they can essentially gamify virality by making an...effect that fits the “Instagram face” aesthetic” “This is something we don't speak about—that deformation can make your filter go viral” https://www.technologyreview.com/ ... Allison Arieff / @aarieff : will never ever ever understand Instagram face https://www.technologyreview.com/ ... @hypervisible : “The little research that exists about digital beauty culture has found that visual platforms like Instagram, which rely on AI recommendation algorithms, are narrowing beauty standards at a stunningly rapid pace.” https://www.technologyreview.com/ ...
Context & Ripple Effects
Meta's October 2021 rebrand put augmented reality at the center of the company's future, but the AR it already ships — face filters on Instagram — has been under scrutiny since MIT Technology Review's 2021 investigation into how beauty filters affect girls and young women. This follow-up finds the core tension unresolved: Meta has banned some deformation filters, yet creator and user demand keeps them viral.
The stakes extend beyond Instagram. Experts flagged in Meta's metaverse avatar plans worry personalized digital bodies could deepen body-image problems, meaning the filter debate is a preview of the identity questions Meta's pivot creates at larger scale.
First-order effects
- Filter creators face a split incentive: Meta's bans remove specific effects from Spark AR, but consistent demand lets creators who keep shipping within the rules effectively gamify virality, as Matt Navarra notes.
- Meta must enforce against the very effect category its own platforms reward, putting its trust-and-safety team in direct conflict with its engagement machinery.
Second-order effects
- As Meta pushes TikTok-style unconnected recommendations per its internal memo, algorithmic distribution amplifies whatever filters perform — including the deformation effects it has pledged to curb — raising the enforcement load.
- Rival visual platforms inherit the same dilemma: banning harmful filters costs engagement, keeping them invites the regulatory and reputational exposure Meta is already absorbing.
Third-order effects
- If the pattern holds, synthetic self-representation becomes a standing governance problem rather than a one-off moderation case — a problem that scales up as Meta automates identity further through features like AI-generated images of users and AI characters populating its apps.
- Sustained evidence of harm to young users positions beauty filters and their successors as likely targets for platform regulation, forcing Meta to build moderation infrastructure for generated appearance itself.
The trend: Platforms are shifting from moderating individual harmful effects to governing synthetic self-representation wholesale, as filters give way to fully AI-generated identities across Meta's apps.