Mosseri is right about the AI backlash, but forecasting it isn't a plan

When the person running one of the world's biggest feeds tells you people are going to get angrier about AI, is that a warning or an alibi? Both, probably. And the second reading is the one worth dwelling on.

Here is what we know. As The Verge reports, Instagram head Adam Mosseri said in an interview that he expects frustration and anger toward AI to keep getting worse, and not only in creative industries. Explainx notes that his list of worries covers training on creators' work, deepfakes and jobs. Other coverage says he made the remarks on the Colin and Samir podcast, mainly while discussing Instagram's new widescreen and long-form video launch.

He's right, and the evidence is already outside the feed

We don't need to take Mosseri's word for it. The anger is showing up in places that have nothing to do with Instagram. Channel NewsAsia reports that China's Lao Gan Ma denied mass layoffs linked to AI amid public backlash. In Mississippi, according to a Guardian investigation and Mississippi Today reporting relayed by Italian outlet Online-News, residents near Elon Musk's AI data centers complain about noise and vibration, and some are bristling at purchase offers that ask them to give up joining a lawsuit.

Different stories, same pattern: AI stops being an abstraction the moment it lands on someone's job, sleep or timeline. Our view is that the backlash isn't irrational technophobia. It is people reacting to costs they didn't choose, while the benefits are described in keynotes. Mosseri's prediction is less a hot take than a weather report.

Predicting anger is the easy part

Here is where we part ways with the tone of the coverage. A platform executive forecasting public fury over deepfakes and AI content is describing a problem his own industry helped create, in our view. Feeds were the distribution channel. Ranking systems decided what got seen. If trust is eroding, and reports of Mosseri's comments say the flood of AI images is wearing down trust in Instagram while users start valuing raw, imperfect content as proof of authenticity, then the company holding the megaphone has a duty that goes beyond commentary.

Mosseri himself has made a candid point, according to MSN's summary: it is becoming impossible to tell real photos from fake ones because of AI. We agree. But that admission cuts against the idea that the burden should sit with users squinting at pixels. If you can't tell by looking, the platform has to tell you. That means labels that are visible, consistent and hard to dodge, plus clear information about who made what. Reports say Meta has been rolling out new labels and creator information. The question is whether they are strong enough to survive contact with people who want to deceive.

There is also a fairness point. Mosseri reportedly notes that people enjoy complaining about AI slop while a lot of AI-made content is great. That's a fair point, and nobody serious wants to ban the good stuff. But "some of it is great" is exactly why sorting matters. Good AI content loses nothing from an honest label. Only the deceptive kind does.

"Control" has to mean more than a smarter algorithm

The most telling line of the discussion, as Italian site iGizmo frames it, is the idea that if you don't like AI, you shouldn't have it in your feed. Sounds reasonable. The catch, as that outlet points out, is that leaving the selection to the algorithm gives users indirect control, not absolute control. We'd go further: a ranking system that guesses your preferences from behavior is not a setting. It is a suggestion you can't audit.

Real control would look boring and concrete. A toggle that reduces AI-generated content. Labels you can filter on. Likeness protection that works for people who never asked to be in a synthetic video. Explainx's coverage mentions platform likeness detection and Facebook's "Your Algorithm" feed controls, which suggests the tools aren't science fiction. What's missing, in our opinion, is making them the default rather than a menu item you find after a scandal.

Our verdict

Take Mosseri's warning seriously, because he is probably right. But don't let it pass as leadership. A forecast of backlash is what you say when you want credit for foresight without taking responsibility for the cause. If Instagram and the rest of Meta want the anger to level off, the sequence is simple: label aggressively, protect people's likeness, and hand users genuine switches instead of probabilistic nudges.

For readers, the practical advice is less glamorous. Treat unlabeled, too-perfect images with suspicion, favor creators who show their process, and use whatever feed controls exist today. And for the platforms: the backlash isn't coming. It has arrived. What remains to be decided is whether they get ahead of it or just keep narrating it.