Science YouTuber Hank Green admitted his AI habit spiraled out of control, telling fans the dopamine hit from leaning on chatbots "is not healthy for me or good for the world."
One phrase gave it away. In a recent video on Complexly, the educational studio Green runs with his brother John, he said "I appreciate the pushback" in a spot where no human writer would. Fans caught it fast. Within a day the science communicator behind SciShow, Crash Course and his own Hankschannel was explaining himself in a since-deleted post on X.
He admitted he wrote the video "under a ton of pressure" and leaned on ChatGPT "for research on this script." That wasn't the end of it. According to TechCrunch, Green followed with a longer, more personal account on Reddit. There he called himself "mortified" and said he had "let so many people down."
Here's the line that mattered. Green wrote that he needs to "come to terms with the fact that the level of dopamine that I've been getting from interacting with LLMs...with doing more and more and more and more...is not healthy for me or good for the world." That's not a hedge. It's a confession from someone who's spent nearly two decades building an audience on being straight with them.
Here's what makes this stick beyond one bad script. Green isn't a random creator dabbling in a new tool. He co-founded Complexly, the studio behind SciShow and Crash Course, channels that have taught millions of viewers how to weigh evidence and check sources. He built his career, and a chunk of the internet's science literacy along with it, on being a credible narrator. When that narrator says he can't tell anymore whether he's leaning on AI too much, it lands differently than a lifestyle influencer admitting the same thing.
The Family Noticed First #
Green wasn't the only one who'd noticed. According to Gizmodo, his wife Katherine and his brother John had already raised concerns about his AI use before the script controversy went public. The viewer backlash just forced that conversation into the open. Green told fans he isn't "a pure AI-hater." He still listed real objections: the climate cost of running these models, and what he called the speed at which AI companies are trying to consolidate economic power.
The fix he's proposing is concrete, not aspirational. Per reporting from Kotaku and TechCrunch, Green says he's pausing two projects, SMUSH and 4x3. Hankschannel itself may go quiet for a while, he says, as he redirects his attention toward a longer-term project and resets how he works with the tools he'd been using daily.
Nobody made him say any of this.
A Lesson for Builders #
That's the part worth sitting with if you build products, run a team, or just spend your workday bouncing ideas off a chatbot. Green's account reads less like a scandal and more like a symptom. He describes a loop familiar to anyone who's caught themselves reaching for ChatGPT before their own brain: the tool works, the output gets praise. So you use it more. Then the muscle you built your reputation on starts to atrophy. He didn't call it addiction in a clinical sense. But he used the word "dopamine" for a reason, and it's the same word behavioral researchers reach for to describe slot machines, not spreadsheets.
Founders building AI products love to talk about engagement and retention. Green's apology is a reminder. Those numbers can climb for the wrong reasons. A creator getting so much value from a tool that it starts eroding the credibility his business depends on is a data point every product team chasing daily-active-use metrics should sit with.
For now, Green's plan is smaller output and closer scrutiny of his own habits, not a wholesale rejection of AI. That's the trade he's making. Whether his audience, and the wider creator economy watching this unfold, treats that as enough is another question. It will say a lot about how forgiving people are willing to be with public figures who admit the tools got the better of them before anyone forced the issue. Also read: Iran-Linked Hackers Hit 30 Minnesota Water Systems Days After US Warning • Codex Users Are Losing Banked Rate Limit Resets To A Quiet 30 Day Clock • Judge Denies Elon Musk's xAI Bid to Block Minnesota's AI Nudify Ban