‘Not healthy’ LLM use is more common than you think
The Verge Robert Hart ● Covered by 2 sources
Hank Green admits his AI habit for research got "not healthy," so he's stepping back from making videos. He swears he never let it write scripts, but fans are still asking how that squares with his trust-built brand.
Hank Green built a career on being the guy who explains complicated things clearly and honestly, so when he announced he's pulling back from production because his AI use had become "not healthy," the internet did what it does: it piled on. Green was quick to clarify he wasn't using ChatGPT or similar tools to write his scripts. He says he leaned on AI mainly to dig up research sources, the unglamorous grunt work behind every explainer video he's made famous.
That clarification hasn't satisfied everyone. The backlash isn't really about whether Green typed a prompt and got a finished script back. It's about the deeper tension between a creator whose entire audience relationship rests on credibility and a technology that, by design, can hallucinate facts with total confidence. Large language models don't know things the way a human researcher does. They predict plausible-sounding text, and sometimes that text is simply wrong. For someone whose job is being right, or at least defensible, that's not a small caveat.
What's gotten less attention, and probably deserves more, is Green's own admission about how his usage felt from the inside. Calling it "not healthy" is a strange phrase for someone describing a research tool, and it hints at something beyond simple efficiency-seeking. People don't usually describe spreadsheet use or library visits in the language of compulsion. That word choice suggests Green stumbled into a pattern a lot of professionals are quietly noticing in themselves: leaning on AI not because it's better, but because it's easier, and then leaning on it more, and then wondering where the line was.
Green's decision to step back is notable precisely because he's not a fringe figure hawking crypto or chasing outrage clicks. He's mainstream, trusted, and has spent over a decade cultivating an audience that believes he does the homework himself. If someone with that much reputational skin in the game is confessing to an unhealthy relationship with AI tools, it's worth wondering how many other creators, journalists, and knowledge workers are quietly in the same spot, just without the audience paying close enough attention to notice.
My take
Nobody wants to admit they've outsourced their thinking to a chatbot, and Green deserves some credit for saying the quiet part out loud instead of just quietly fixing it behind the scenes. The bigger story here isn't one YouTuber's habits, it's that "AI use" has quietly become a spectrum from tool to crutch to dependency, and most people using it daily have no idea which end they're on. Companies selling these tools have zero incentive to help anyone find that line, so expect more of these public reckonings, not fewer.
Read more about this at: The Verge
