AI-generated filler is crowding social feeds, and a lot of creators are wondering whether their watch time will survive the flood. Science YouTuber Mark Rober has a blunt answer: get off the green screen and do something in the real world.

According to Business Insider‘s Dan Whateley, Rober — who has more than 85 million YouTube subscribers and founded the STEM kit company CrunchLabs — told the outlet that authenticity is the edge machine-made clips still struggle to fake.

“Do you want to see a simulation of someone dunking a basketball with a sick trick, or do you want to see someone actually do it?” he said. “The fact that you know it’s real is what makes it crazy.”

A “moat” built on being there

Rober argued that human connection still matters because people evolved around it for hundreds of thousands of years, and that craving for “real and authentic” work gives on-camera creators a temporary advantage. “At least for a while, I think I have a moat on the YouTube channel,” he told Business Insider.

That moat is not evenly distributed. Whateley notes that AI can speed up editing and help content travel farther through tools like language dubbing, but it also commodifies footage — a tougher hit for faceless or fully animated channels than for creators whose faces and audiences are already recognizable.

Platforms are pushing back, too

The same Business Insider piece maps how major platforms are starting to treat AI filler as a product problem. YouTube is demonetizing AI content that looks mass-produced or generic under its inauthentic content policy. LinkedIn lets users flag posts that “seem like AI slop.” TikTok offers a toggle to cut how much AI content appears in the feed, and Substack has a Pangram-powered scanner for AI writing in newsletters.

Using AI as a tool, not the whole show

According to Tubefilter‘s James Hale, Rober’s own shop is not anti-AI. His team uses generative AI as a “thought partner” for writing and some thumbnail work, and he has described Class CrunchLabs — a free science curriculum for third- through eighth-graders, backed as a Google/YouTube partner with a $2.5 million Google.org grant — as a stack of strong “ingredients” that AI can remix for different classrooms.

In science videos specifically, he still bets audiences want the real thing. “I would think in science, you actually want to see the lava melt through the thing, versus it just being completely fake,” he told Tubefilter. “But we’ll see.”

Hale frames the contradiction many big creators now live with: they feel protected from other people’s AI slop while quietly using AI to scale production themselves — the same catch-22 platforms face when they sell generative tools and then have to clean up the low-quality flood those tools create.

CrunchLabs over the personal brand

Rober told Business Insider his longer game is to put the company ahead of the personality. “I want CrunchLabs to be a lot more well known than Mark Rober,” he said. The company broke even on its YouTube channels last year with about $17 million in revenue matched by similar costs; profits come from CrunchLabs products and kits. It now employs more than 200 people, he said, with a mission centered on curiosity and learning.

“Curiosity is so important in this new age of AI,” Rober said. “It means you just have a lifetime of learning, versus like, ‘Oh, this thing knows the answer, so who cares?'”

The takeaway

Rober’s pitch to the creator economy is simple: AI can help you edit, title, and adapt — but the work that still cuts through is the work no model can stage-manage as well as a person who actually did it. Platforms are starting to punish generic AI filler. Creators who lean harder into real stunts, real faces, and real experiments may keep the “moat” longest. Everyone else is competing with infinite B-roll.

Image: Mark Rober in September 2024, photo by LKP2024, Wikimedia Commons (CC BY 4.0).