Ranting On...Hank Green and the Complicated World of AI

I have made a concerted effort in the past year to genuinely reduce the amount of time I spend on my phone.  This has been a mixed bag.  Partially, it's because for the true time-killers (for me, that's Twitter and waiting for texts) I have transferred to just watching on my laptop, which is still better (because I also do other things like actual hobbies, writing, and work) on those, but still not what I wanted, which was to truly cut back on my wasted online time.  But what it did do (because I have cut well over an hour off of my average screen time compared to, say, December & January) is cause me to look at video websites like TikTok and YouTube much, much less.  In fact, for someone who used to religiously keep up-to-date on YouTube videos, for me it's pretty much just Bad Boy that I consider "appointment viewing" that I will not miss a single episode.  But that doesn't mean that I don't keep abreast with some of my past YouTube obsessions, and that includes Hank Green, one of the two hosts of The Vlog Brothers, the extremely long-running YouTube channel that launched both Hank and his brother John into an echelon of pedigreed YouTube stars and of whichI have been a viewer for many, many years (looking through the page, I have seen several in 2026 even with my cutback on screen-time).  John is a New York Times bestselling author whose books The Fault in Our Stars and Paper Towns have both been made into major motion-pictures, while Hank has become an essential part of America's classrooms, with his series like Crash Course and SciShow regularly shown by teachers in educating their students.  For a certain generation, Hank Green is a hero, and though I'm a bit too old to be calling him that (even though it wouldn't entirely be untrue), I'm definitely a fan.  I've seen his show with his brother live, and I have a framed photo of me meeting both of them on the dresser in my bedroom.

Which is why when there was a fan revolt over Green in the past two weeks, I stood up and took notice, and wanted to get an article out about it even though my To Do list right now is drowning me.  Essentially, in a recent video, Green used the phrase "I appreciate the pushback" which caused fans to wonder if Green had used an AI program to write his script for the "Future Hank" segment of the video.  Anyone who has spent much time on AI knows that "I appreciate the pushback" is a tell that you are using AI to write something, as it's a commonly used phrase by the programming when you are challenging it to create something better.  Green admitted in a Twitter post (that has since been deleted) that he had used ChatGPT for research on the script (which was about how the sounds of words help to create their construction), but did claim that "I appreciate the pushback" was something he said of his own volition.

This caused a significant amount of reaction online, something that seems to follow literally any use of AI in modern culture.  For starters, Green saying he used AI to research his video but not to write it made some viewers/followers doubtful-you can see countless people saying that they don't think he actually was being truthful here, which given that he's using words that sound like AI, well, gives cause for doubt.  But I'm going to take Green at his word, because I've watched enough of his videos (for decades) that that sentence doesn't feel out-of-character for me.  Green has a thoughtful, nerdy, and sometimes sporadic cadence when he speaks (whereas his brother is a bit more contained), and so "I appreciate the pushback" feels at home for him.

But I do think this controversy opens up some other doors.  Green, in an apology on Reddit, pointed out that some of the people who were attacking him seemed to be doing so in large part due to his use of AI at all.  Green acknowledges the environmental and ethical ramifications of the tool, but also is pretty clear that he has a more nuanced view of it, which some people may disagree with.  He then goes on to point out that he has become over-reliant on LLM's and AI in his research, which has impacted his relationships not just to his work, but to the people around him (he name checks both his wife Katherine and his brother John as agreeing with the latter statement, which makes it feel a little bit more like a family issue than one involving his fans).

This opens up two questions for me, so let's start with the first one.  I have written a couple of articles on this blog about AI and ChatGPT specifically.  I struggle with something like ChatGPT for a variety of reasons, the biggest being the environmental impact.  I also think, to a degree the "either use it or get left behind" aspect is a bit self-fulfilling prophecy (we're only being left behind because we're being forced to be left behind), but I will own that I have used AI and ChatGPT specifically, both professionally (this was heavily encouraged at my place of business) and personally, which I recounted here (I will note, as I have in pretty much every mention of the technology on this article in the past, that I will never use it for writing a blog article, as honestly I'd rather shut the place down than use AI to write an article on here-everything on here, including the mistakes, is all me).  I think, ultimately, the environmental aspects of it are hard to fault Green for, because literally everyone's using AI (just try googling without using it), so why shouldn't he too?  But the academic implications of it, particularly given how much of AI is inaccurate (which I also profiled in that article if you haven't already read it-it's a good one), is where Green gets into trouble.  Green has become a trusted source on the internet (people literally tag him in TikTok videos to see if he can explain phenomena in videos); that he's using a tool that is rife with inaccuracy (fairly or unfairly) makes people doubt the veracity of his videos.

I also think that what he says about ChatGPT being addictive (we're coming to the end, so we're bringing this back to the first paragraph about screen time...proof there's a method to my madness) is really important.  ChatGPT I have used in my personal life a few times, and it is shockingly addictive.  Partially, to Hank's credit, it's voluminous as a research tool.  You need to always double check everything if you're using it, which requires extra work that some won't put in, but if you ask it for, say, a list of movie musicals made in 1950, it's going to give you a better list than if you Google that same prompt (this is partially because Google has denigrated its quality in the past five years, a conversation for a different day).  It's also a decent brainstorming tool-I recently was working with ChatGPT to plan a trip and trying to get all of the things on my wishlist for it done, and eventually through a lot of pushback (forbidden word choice use intentional), I realized that the trip I was planning was impossible the way I was planning it, and came up with a Plan B I probably wouldn't have realized in the same amount of time on my own, and transparently maybe not at all.

But it also shapes the way you research, sometimes to your detriment.  I can type in a prompt asking for a list of all of the movie musicals of 1950, but it also gives opinions on those movies, some of which are gleaned from less-than-reputable sites where those opinions are presented as facts (and ChatGPT interprets them as facts).  This is dangerous because the human brain is not meant to constantly sort through things like that, but also because it's hard not to assume that the research & direction of that approach aren't where you need to go.  ChatGPT gave you an idea...even if it's a good idea, it's still ChatGPT's idea.  You didn't do that thinking, you are using the AI algorithm to create that idea and it is informing your research.  And because it's hard work to create ideas, having something do that thinking for you, particularly with effusive praise and lots of encouragement as if it were your idea is scary, and can make you think you can do more than you can.  This seems to be what Green is getting after-this is causing him to over-extend, and the more you over-extend, the more you think you need ChatGPT to work.  It's hard not to read those sentences, and replace it with a word like, say, alcohol or gambling, and not realize that there's an under-discussed addictive quality to AI/LLM's that we aren't allowing for in understanding its value.  That, more than anything, feels not like a scandal (I remain pro-Hank Green after this, and maybe even more so because of the honesty he brought forth in showing how challenging using this tool can be), but a wake up call.  If Hank Green, endlessly creative & boundlessly trying to make the world a better place, can fall prey to ChatGPT...anyone can.