Instagram won't ban AI feeds, but wants to let you dodge them

Instagram's algorithm has been aggressively pushing synthetic media lately, but according to platform chief Adam Mosseri, you shouldn't have to look at it if you hate it.

Speaking on Lenny Rachitsky's podcast, Mosseri weighed in on the growing flood of generative AI on social media. While he insisted that users who dislike AI content should not have it clogging their feeds, he drew a firm line against platform-wide bans or blanket filters.

An 'AI town' feed, but no off switch

Mosseri wants Instagram to focus on labeling rather than censorship. In his view, users who love synthetic creations should be free to live in an all-AI feed, which he called an "AI town."

Yet despite that vision, Instagram still behaves like its peers TikTok, YouTube, and Facebook: it attaches labels to generated media when detected, but offers no actual toggle to mute AI posts entirely.

Catching AI content is only getting harder, Mosseri acknowledged. In fact, as generative models keep evolving, he warned that platforms could eventually lose the ability to detect synthetic posts altogether.

Instead of hunting down AI fakes, Mosseri argued it might be far more realistic to verify "camera-captured content" with hardware fingerprints, echoing an idea he first floated in December 2025. In the future, he imagines users simply asking the app, "Is this AI?" and getting a probabilistic answer ranging from likely to definitely not.

Pushing tools while spam multiplies

Even as Mosseri admitted that Instagram must figure out how to rein in spammy AI posts, Meta is rushing out high-risk generative features.

Instagram recently rolled out Muse Spark, an AI image generator that allows people to drop other users directly into AI creations simply by tagging their handle.

Child safety advocates are already sounding the alarm. Haley McNamara, executive director of the National Center on Sexual Exploitation, warned that the tagging tool creates obvious, foreseeable dangers for harassment, exploitation, identity fraud, and abuse.

Why it matters

For everyday Instagram users, Mosseri's comments feel contradictory. Saying users shouldn't have to see AI without actually giving them an opt-out toggle leaves frustrated creators and casual scrollers stranded in an increasingly synthetic feed.

We think Mosseri's concession that automated AI detection will eventually fail is brutally honest. Pivoting to certifying authenticated camera captures is likely the only long-term technical fix, though it turns organic reality into a premium badge that creators have to verify.

What looks problematic is Meta's aggressive deployment of tools like Muse Spark. Rushing out features that let anyone generate imagery of another person via a simple tag, while simultaneously struggling to police spam, shows where priorities lie. In the race to lock in AI engagement, safety guardrails still look like an afterthought.