Hank Green
Science Communicator & YouTuber
Co-creator of Vlogbrothers, SciShow, and Crash Course, and founder of VidCon. He draws a hard line down the middle of AI: as a utility — code, drafting, drudgery — it is “mostly, weirdly, kind of fine,” and he uses ChatGPT and ships his own AI-built software; as engagement-optimized slop and consent-free training it degrades the media ecosystem he helped build, which is why he savaged Sora 2’s “SlopTok” and demanded creators be able to opt out.
In his words
I feel like I should be ashamed for using ChatGPT, but I’m not sure why. It isn’t good at anything except getting me started, but that’s, like, the thing I am worst at.
There are a lot of critiques of LLMs that I agree with but “they suck and aren’t useful” doesn’t really hold water.
My Bluesky: “Generative AI is just fancy autocomplete.” My YouTube: “Generative AI will deliver century of progress in the next decade.” I’m going to go out on a limb and say our situation somewhere between these two.
Biography
Hank Green (born 1980) is one of the architects of educational YouTube. With his brother John he launched Vlogbrothers in 2007 and the Nerdfighter community around it; his company Complexly produces SciShow and Crash Course, staples of classrooms worldwide; and he co-founded VidCon, the creator economy’s flagship convention. He is also a bestselling novelist and one of the platform era’s most trusted explainers of how the internet actually works.
A user and a builder
Green’s relationship with AI starts with the fact that he uses it. He admitted to using ChatGPT as a drafting aid back in March 2023, told Dear Hank & John listeners in 2025 about asking ChatGPT to build a research spreadsheet that “made it better than I could have made it,” and pushed back on Bluesky against the claim that LLMs “suck and aren’t useful.” He is also shipping software again: his daily word game 4×3 (June 2026) is a solo-built web app he continually patches — “a couple of things that I have coded into this thing,” as he put it while walking through its error handling — and a reviewing front-end engineer noted distinctly Claude-flavored patterns in its code. His Focus Friend focus-timer app, built with developer Bria Sullivan, hit #1 on the App Store in August 2025 — above ChatGPT. In April 2026 he even made a video debunking exaggerated AI water-use statistics, explaining he “was tired of seeing AI critics lose credibility because they were focusing on weak talking points and fake statistics.”
The loudest critic among the users
None of that has blunted his criticism. A July 2024 Proof News investigation found subtitles from SciShow and Crash Course videos in AI training sets from Apple, Nvidia, and Anthropic; Green told his audience he was being ripped off and argued creators should at minimum be able to opt out. When OpenAI launched the Sora 2 social app in October 2025 he called the feed “SlopTok,” a “lawsuit magnet,” and asked for “a single reason why Sora 2 should exist.” His video “The State of the AI Industry is Freaking Me Out” dissected the buildout’s circular economics, and in August 2025 he moved a quarter of his S&P 500 index money into value, mid-cap, and international funds because the index had become “kind of betting on a big AI future.”
I do not care whether a programmer spent half their life making that tool or whether it was created by Claude in an afternoon. I just care how easy the tool is to use and how good it is at its job.
The through-line is a distinction he drew in his March 2026 “What is Slop” video: AI as utility — tools, code, drudgery elimination — is “mostly, weirdly, kind of fine,” while AI as an engagement-optimized synthetic feed degrades the media ecosystem he helped build. Unlike creators such as Casey Neistat he has filed no lawsuit and signed no protest letter; like MrBeast, he complains loudly about where the industry is heading while building with the tools himself.
Where they stand in the war
Voices for AI
Sources & further reading
Canonical record: https://battlelines.ai/topic/hank-green