Instagram chief Adam Mosseri said AI will change who can be creative, as new tools and technologies will allow people who previously couldn’t be creators to create content at a certain quality and scale. But he also acknowledged that bad actors will use the technology for “illicit purposes” and that children growing up today will have to be taught not to believe something just because they see a video.
Meta executives shared their thoughts on how AI is impacting the creator industry at this week’s Bloomberg Screen Time conference. At the beginning of the interview, Mosseri was asked about recent comments from creator Mr. Beast (Jimmy Donaldson). In the thread, Beast suggested that AI-generated videos could soon threaten creators’ livelihoods, calling it a “scary time” for the industry.
Mosseri pushed back a bit on that idea, pointing out that most creators wouldn’t use AI technology to recreate the huge sets and elaborate productions that Mr. Beast has historically done. In return, creators can do more and create better content.
“If you take a big step back, the Internet’s accomplishment, among other things, is that it has made it possible for almost anyone to become a publisher by reducing the cost of distributing content to virtually zero,” Mosseri explained. “And some of these generative AI models seem to be aiming to reduce the cost of producing content to essentially zero,” he said. (Of course, this does not reflect the real economic, environmental, and human costs of using AI; these costs are substantial.)
The executive also suggested that today’s big social platforms already have a lot of “hybrid” content, with creators using AI in their workflows, but not producing fully synthetic content. For example, you may be using AI tools for color correction or filters. Mosseri said that in the future, the line between what is real and what is generated by AI will become even more blurred.
“The question of what is organic content, what is AI-synthesized content, and what proportions of content is going to fade away little by little. I think for a while there will actually be more content in between than pure synthetic content,” he said.
As the landscape changes, Meta has a responsibility to do more in terms of identifying AI-generated content, Mosseri said. But he also pointed out that the company’s previous approach was not the “right focus” and was effectively a “fool’s errand.” He was referring to how Meta initially tried to automatically label AI content and ended up in a situation where it labeled actual content as AI because AI tools like Adobe were used as part of the process.
tech crunch event
san francisco
|
October 27-29, 2025
The executive said more work needs to be done on the labeling system, but Meta needs to provide more context so people can make informed decisions.
He didn’t elaborate on what the new context would be, but he may have been thinking about Meta’s community notes feature. This is a crowd-sourced fact-checking system launched in the United States this year, modeled after the system used by Company X. Rather than relying on third-party fact checkers, community notes and similar systems mark content with corrections or additional context when users who often share opposing views agree that it requires fact checking or further clarification. Perhaps Meta could be considering using such a system to flag when something is generated by AI but not labeled as such.
Rather than saying that labeling AI content is entirely the responsibility of platforms, Mosseri suggested that society itself will have to change.
“My kids are young. They’re 9, 7, and 5. When they grow up and have access to the internet, I need them to understand that just because they’re watching a video of something doesn’t mean it actually happened,” he explained. “When I grew up and watched the video, I could see it as a capture of a moment that happened in the real world,” Mosseri continued.
“What they’re trying to do… in this case, you need to think about who’s saying it, who’s sharing it, what their incentives are, why they’re saying it,” he concluded. (That seems like a big mental burden for a 5- to 9-year-old child, unfortunately.)
During the discussion, Mosseri touched on other topics about Instagram’s future beyond AI, including plans for a dedicated TV app, a new focus on Reels and DMs as core features (which Mosseri said simply reflects user trends), and how TikTok’s ownership change in the U.S. will impact the competitive landscape.
On the latter, he said, ultimately competition is better, as TikTok’s U.S. presence forces Instagram to “do a better job.” As for TikTok’s deal itself, Mosseri said it’s difficult to parse, but it doesn’t appear to change the way the app is built in any meaningful way.
“Same app, same ranking system, same creators you follow – same people. Everything is seamless,” Mosseri said of the “new” TikTok US operation. “It doesn’t seem like a big change in terms of incentives,” he added.
