Hank Green, who runs the educational-media company Complexly behind SciShow and Crash Course, apologised over the weekend for using ChatGPT while researching a script — after viewers spotted the tell themselves.
How he was caught
No classifier and no platform label were involved. The audience flagged a phrase in the video, "I appreciate the pushback," as reading like model output, and the complaint spread from there. Green then posted on X, since deleted, and on Reddit.
What he said he would change
Green said he had used ChatGPT for script research and that he would cut his upload frequency. On the wider habit, TechCrunch quotes him directly: the dopamine he has been getting from interacting with LLMs "is not healthy for me or good for the world."
Two versions of the same admission
The outlets that covered it do not tell the same story. TechCrunch attributes that health judgement to Green. Gizmodo frames it as the view of people close to him — that his loved ones do not think his LLM use has been healthy. Those are materially different admissions, and the primary record is thin: the X post was deleted, and the Reddit thread is the only first-hand version left. Merging the two accounts into one sentence, which several rewrites have already done, produces a quote nobody gave.
Why an education channel is the awkward place for this
Complexly's product is explanatory science content, where the implicit promise is that a person checked the claim. Green's channel is cited at 3.2 million subscribers. The enforcement ran bottom-up: no disclosure policy, no detector, no platform intervention — just an audience that reads enough of one writer's work to notice when the cadence changes.