Hank Green apologises for using ChatGPT to help write his YouTube scripts
The popular science communicator says he got caught up in the dopamine hit of doing more with AI, and admits he lost track of how much the technology shaped his own words.
Key points
- Hank Green, a YouTuber with 3.2 million subscribers, apologised in May 2025 for using ChatGPT to help research and write scripts for his educational videos.
- Viewers suspected AI involvement after Green accidentally left in the phrase "I appreciate the pushback," a turn of phrase typical of chatbot responses.
- Green deleted his initial explanation on X and posted a longer apology on Reddit, calling himself "mortified."
- He plans to slow down his output and return to writing he "felt all the way down."
- Green listed climate costs and corporate consolidation among his concerns about AI, even as he admitted he had been using it more than was healthy.
Hank Green makes science feel approachable. Across several YouTube channels, including the educational network Complexly, his 3.2 million subscribers tune in for his own voice. So when that voice started sounding a little off, people noticed.
The trouble began mid-video, when Green used the phrase "I appreciate the pushback." It's a line chatbots produce constantly. Viewers flagged it immediately, wondering whether it had slipped in from a ChatGPT response rather than from a real conversation with the episode's guest. According to SOURCE, Green later clarified the line was actually a response to that guest, not leftover chatbot text.
What did Hank Green actually admit?
He confirmed using ChatGPT to help research scripts, while insisting the opinions and final wording were his own. He acknowledged the criticism was fair.
Green first addressed the story on X, then deleted the post. His fuller Reddit apology described feeling "mortified" for having "let so many people down." He framed his AI use as a research shortcut, locating academic papers and background reading rather than generating finished sentences, but admitted it was reasonable to call that "diluting" himself.
First reported by TechCrunch AI, the story spread quickly because Green is one of the more prominent voices in science communication, a space where trust in the author matters enormously. We covered a related moment when a Canadian politician accidentally read his AI chatbot's instructions aloud during a floor speech, and the lesson there was the same: it's surprisingly easy to lose track of where your thinking ends and the machine's begins.
Why does this matter for ordinary viewers?
When you watch an educational video, you're trusting that the person talking has genuinely worked through the material. If an AI summarised the research, the presenter may have absorbed a simplified or subtly wrong version without realising it.
Green made a version of this point himself. "I've been moving so fast that my own process isn't actually clear to me," he wrote. Speed, helped along by AI, had blurred the line between his thinking and the machine's.
He also called out the addictive pull directly. "The level of dopamine I've been getting from interacting with LLMs," he wrote, using the shorthand for large language models, the AI systems behind tools like ChatGPT, "is not healthy for me or good for the world."
What happens next?
Green is posting less. He pointed to a recent meditative video as the kind of work he wants to do more of, one "where the writing was the whole thing." More "dumb unscripted straight to camera thoughts" are probably coming too, which is about as clear a signal as a creator can send that they're stepping back from AI assistance.
Green says he's not an AI opponent. His listed concerns are real: the energy AI data centres consume and the speed at which big tech companies are trying to consolidate economic power. But his main worry sounds more personal. "What I am most scared of," he wrote, "is ruining myself for people."
