Month: August 2026

So Hank Green got caught using AI to help write a script, and the internet, predictably, lost its collective mind. Within a day there were think pieces, quote tweets, Reddit threads with thousands of comments, all of it circling the same basic accusation: Hank Green, patron saint of curious nerds everywhere, co-founder of Crash Course and SciShow and basically every YouTube channel that made learning feel like fun instead of homework, had used ChatGPT. Gasp. The horror. Bring out the pitchforks, roll out the tumbrels, let’s parade him through the digital town square. And look, I get why people care. I do. But I also think a huge chunk of this reaction is theater, and I want to talk about why.

Here’s the thing that keeps getting lost in all of this. As far as anyone has actually established, Hank didn’t have an AI write his script. He used it for research, for finding papers, for organizing notes, and by his own admission it bled into his phrasing more than he wanted it to, which is a legitimate thing to be embarrassed about and a legitimate thing to apologize for. But that’s a different crime than the one people are prosecuting him for in the group chat. People are acting like he handed his entire creative process over to a chatbot and pressed generate, like he’s some hollowed-out puppet reciting whatever the machine spat out, and that’s just not what happened. He was still there. He was still doing the thinking, still doing the talking, still the one standing in front of the camera taking responsibility for what got said. The AI helped him gather string faster. That’s it. That’s the scandal.

And even if the scandal were bigger than that, even if he’d used AI more heavily to draft actual sentences that ended up in the video, I keep coming back to the same question: so what? Like, genuinely, what is the actual harm being alleged here? Because when I look at the arguments people are making, they tend to fall into one of a few categories, and none of them hold up as well as the outrage volume would suggest.

The first argument is some version of “it’s cheating.” As if writing a YouTube video is a proctored exam and there’s a rulebook somewhere that says you’re only allowed to use your own brain, alone, in a locked room, with no tools. But nobody actually believes this rule for anything else. Nobody says it’s cheating to use a thesaurus, or a citation manager, or an editor, or a research assistant, or Google, or a stack of library books someone else summarized for you. Writers have always used tools to think faster and organize better. The printing press was a tool. Spellcheck is a tool. An AI that helps you find and synthesize research faster is a tool. If the end result is something a person stands behind, edits, performs, and takes ownership of, then the fact that a tool sped up part of the process isn’t a moral failing, it’s just how work gets done in every field that isn’t monk-copying manuscripts by candlelight.

The second argument is about authenticity, this idea that when we watch Hank Green we’re paying for Hank Green’s actual voice, unmediated, and if a machine touched any part of that process then we’ve been sold a counterfeit. I understand the emotional logic here even if I don’t buy the conclusion. Nerdfighteria, the whole community around Hank and John Green, has always been built on this feeling of intimacy, like these are two guys who actually give a shit about you and about ideas and about making the world suck less. When people feel like that trust has been violated, of course they’re going to react hard. But intimacy and authenticity were never about the literal mechanical process of how a script got assembled. They were about whether the person on screen actually believes what they’re saying and actually cares about the audience. Hank clearly still does. His apology wasn’t corporate, it wasn’t a PR statement written by a crisis management firm, it was him going on Reddit and saying he felt mortified and that the dopamine hit of working with LLMs had become unhealthy for him. That’s not the voice of someone who’s outsourced his soul to a chatbot. That’s a guy being honest about a bad habit in real time, in public, in a way that’s actually kind of brave given how brutal the response was going to be.

The third argument, and probably the most serious one, is about misinformation and trust in educational content specifically. This one I take more seriously than the others, because Complexly’s entire brand is built on being a place where people can learn real things and trust that the facts have been checked. If AI-generated research introduced errors into a science video, that’s a real problem, not a vibes problem. That’s the kind of thing that actually matters, because kids and curious adults use these channels as a starting point for understanding the world, and if the pipeline that produces that content gets contaminated with confidently wrong AI output, that erodes something valuable. But notice that this is a different complaint than “he used AI,” full stop. This is a complaint about quality control, about whether the facts in the video were accurate, about whether someone verified the output before it went out into the world. That’s a real conversation to have. It’s just not the conversation most people were actually having. Most people weren’t asking “were the claims in this video accurate,” they were performing disgust at the mere presence of the tool, treating “used AI” as an inherently disqualifying act regardless of what it produced.

And that’s the distinction I keep wanting people to make and keep watching them fail to make. The tool isn’t the sin. The tool is neutral. What matters is what you do with it. If you use AI to fabricate quotes, to invent sources, to generate misleading information and pass it off as verified fact, that’s bad, and it’s bad because of the fabrication and the deception, not because a language model was involved somewhere upstream. If you use AI to steal someone’s specific creative voice or lift copyrighted material wholesale, that’s bad too, and it’s bad because of the theft, not because a computer helped. But if you use AI to speed up research, to help you organize your thoughts, to catch typos, to brainstorm structure, and then a human being still does the actual work of deciding what’s true, what’s worth saying, and how to say it in their own voice, I don’t see the crime. I see someone using a calculator instead of doing long division by hand. The math still has to be right. The person still has to check it. But nobody thinks you’re a fraud for using the calculator.

I think part of what’s actually going on here, underneath all the specific arguments about Hank Green, is a much bigger anxiety about AI in general, and Hank got caught in the blast radius of that anxiety because he’s a visible, beloved, trusted figure who represents a kind of old-school internet sincerity that people are terrified of losing. There’s a real fear running through a lot of online culture right now that AI is going to hollow out human creativity, flood every platform with slop, and make it impossible to tell what’s real anymore. That fear is not irrational. It’s actually pretty well-founded if you look at what’s happened to search results, to social media feeds, to the flood of AI-generated books and articles and images clogging up the internet with garbage nobody asked for. People are right to be anxious about that. I share some of that anxiety myself. The economics of AI content generation are genuinely troubling, because it becomes cheaper and cheaper to produce plausible-sounding junk at scale, and that junk competes for attention with people who are actually trying to say something true and original. There’s also a real conversation to be had about the environmental cost of running these massive models, about the energy and water consumption, about whether the tradeoffs are worth it for every trivial use case people put them to. There’s a conversation to be had about labor, about writers and researchers and assistants who used to get paid to do the kind of work that AI now does for pennies, and about what happens to those jobs and those people as the tools get better. There’s a conversation to be had about the way constant reliance on these tools can atrophy your own skills, can make you lazier, can create exactly the kind of unhealthy dependency Hank himself described feeling. All of that is real. None of that is nothing.

But here’s my problem with dumping all of that anxiety onto one guy who used ChatGPT to help find some research papers for a YouTube episode. When you take a legitimate, complicated, structural set of concerns about an entire technology and its economic and cultural effects, and you compress all of that into a personal morality play about one individual’s specific choices, you end up with an outrage that’s disproportionate to the actual offense and also weirdly useless for addressing the actual problem. Hank Green apologizing and pausing his channel does nothing to fix the environmental cost of AI training runs. It does nothing to fix the economics that are squeezing writers out of jobs. It does nothing to slow the flood of low-effort AI slop across the internet, because the people producing that slop were never going to feel shame about it in the first place, that’s kind of the whole point of slop, it’s made by people who don’t care. What public shaming of Hank Green does accomplish is scratching an itch, letting people feel like they did something, like they held someone accountable, when really what happened is a guy who cares a lot about honesty and trust got dragged through the mud for using a tool in a way that, by his own admission, could have been more careful, but wasn’t remotely as bad as the pile-on suggested.

And I think there’s something almost a little cruel about how fast people were to assume the worst possible version of events, to skip past “let me find out what actually happened” and go straight to “let me post my outrage,” because that’s just become the default mode of being online now. Someone gets a whiff of a controversy, and instead of waiting for details, instead of extending the benefit of the doubt to someone who has spent fifteen years building a body of work premised on being thoughtful and honest, people just assume the most damning interpretation and run with it, because outrage is fun, because dunking on someone feels good, because there’s a kind of social currency in being the first person to call something out. Hank Green has spent years talking publicly and critically about the dangers of AI, about low-effort machine generated content flooding social platforms, about the importance of original human perspective. The guy has been on the record as a skeptic. And the second there was a whiff of hypocrisy, people didn’t pause to consider nuance, they went straight for the pile-on, because hypocrisy stories are the most satisfying stories on the internet, they combine outrage with a little dash of schadenfreude, and that combination is basically internet crack.

What I actually find kind of moving about how Hank responded is that he didn’t do the thing public figures usually do when they get caught in a controversy, which is to minimize, deflect, lawyer up their language, issue a statement that says nothing while sounding like it says everything. He went on Reddit, in the actual community built around his work, and told people something uncomfortably honest, that using these tools had started to feel like a bad habit, that the dopamine hit of getting instant help from an LLM had become something he needed to examine, that he was overcommitted and under pressure and had let that pressure erode some of his own standards. That’s not a corporate crisis response. That’s a person being self-critical in public, which is rare, and which is exactly the kind of behavior that should be rewarded with grace instead of a pile-on, because if the lesson people take away from this is “never admit to using AI, deny everything, get better at hiding it,” then all we’ve done is push the actual usage further underground while making public honesty about it more costly. That seems like a worse outcome for everyone, including the people who are worried about transparency and trust in the first place.

None of this means I think AI is some unambiguous good that deserves a free pass on every front. I really don’t think that. I think the concerns about labor displacement are serious and deserve real policy attention, not just individual guilt trips. I think the concerns about misinformation are serious, especially in contexts like education and journalism where accuracy actually matters and where the incentive to move fast can outpace the diligence required to check facts. I think the concerns about environmental cost are serious and worth demanding more transparency about from the companies building these systems. I think there’s something worth mourning in a world where more and more of our written and visual culture gets produced by tools instead of people, even when those tools are used responsibly, because there’s a texture and a specificity to human effort that gets lost at scale even when the individual outputs are fine. I’m not trying to wave all of that away. I’m trying to say that the specific case of Hank Green using AI to help with research on one script, then owning up to it and reflecting on his own habits in public, is not the hill to build all of that anxiety on. It’s a small, human, correctable moment being treated like a betrayal, and the intensity of that treatment says more about how primed people are to feel betrayed by AI right now than it does about anything Hank Green actually did.

At the end of the day, the standard I keep coming back to is pretty simple. Was a human still meaningfully involved in deciding what got said and why. Was the information presented actually true, or did someone let a hallucination slide through unchecked. Was anyone’s work stolen or misrepresented along the way. If the answers to those questions are fine, then the presence of AI somewhere in the workflow isn’t the scandal people want it to be. It’s just a tool, used by a person, who is still accountable for what they put into the world, the same way they’d be accountable if they’d used a research assistant or an editor or a stack of borrowed books. Hold people accountable for lying, for stealing, for spreading things that aren’t true. Don’t hold them accountable for using a calculator to speed up the boring parts of a job that still requires a human brain to finish. That’s the line. Hank Green landing on the wrong side of some smaller, more personal question about workload and habits and burnout is a much more sympathetic and much more human story than the one everyone rushed to tell, and I think it deserves to be treated that way.