rage against the chatbot: why anger alone can't slow down the a.i. hype train

Generative AI is being sold with utopian fantasies. Making its adoption more ethical, reasonable, and less harmful requires cold, unflinching logic.

It appears that I accidentally ventured into the furious rumbling of the Butlerian Jihad that’s been simmering on Threads and parts of TikTok with a detailed post and video arguing that just using AI tools doesn’t make you dumber. It doesn’t mean we should just jump into using these tools with wild abandon, of course. It just means we need to be mindful when we do and consider the context.

But because I said the viral anti-AI narrative of the day was not actually supported by the study in question and it wasn’t peer reviewed, that could’ve only meant one thing. I’m obviously a brain-dead AI shill who doesn’t understand how he’s being exploited by Big Tech and their insane dreams for the future.

But you know what? That’s okay. I am a very part time, niche, very small “creator” and don’t labor under the delusion that strangers online should know that I’ve been talking about these very problems for the last 16 years or so, almost my entire academic and professional life. My reach is also rather limited so, yes, they don’t know who I am, nor do they care, and they’re not going to look at my bio and research me.

They were looking at what they thought was pro-generative AI content questioning its toll, and exploded with rage. One person even used the phrase “tech chud fluffer” in a thread that argued for ignoring any nuance on the issue and that if I disagree with that approach, I can get fucked. Which no, the poster didn’t mean in the fun, literal way, as inferred from the rather strong and obvious context clues.

Again, I get it. It’s an extremely common and rapidly growing sentiment because when a layperson hears something about AI, they immediately think of chatbots and art barf robots that spam their social media feeds with low effort engagement bait slop which, ironically, undermines their future potential. They also think of their bosses and all the oligarchs and CEOs gleefully excited about replacing them with a ChatGPT or Claude subscription with all the subtlety of a drunk bear pawing at a beehive.

People are so upset about these tools and how they’re being used because the tone-deaf, unhinged, and you’re-doomed-peasants discourse by delusional tech bros has completely poisoned and potentially sabotaged what are, at their core, revolutionary inventions being all too frequently used back asswards.

Meanwhile, when experts in comp sci disciplines hear the term, the first thing we ask is what’s the AI in question. There are hundreds of different ML models and dozens of specific architectures for different goals. That’s not to mention the broader subject of machine learning where all those AI models are a part of much larger questions.

This distinction does matter because there’s a lot more to AI than chatbots, slop, and creepy images and videos on your feeds. There are also weather predictions, ways to defend against chemical weapons, help in fighting genetic diseases, discovery of new antibiotics, improvement in cancer detection and treatment, and so on, and so forth.

So, the way I look at it, asking if one is pro or anti-AI is kind of like asking someone if they’re pro or anti-fork. Forks are great. If someone needs a utensil for meals, I would highly recommend it. What I would not recommend though, is to use the fork to stab people, or shove it into electrical outlets, and that’s all too often what we’re seeing an awful lot of forks being deployed to do right now.

But since AIs are already released, they’re not going to be un-invented. Since they’re already trained, the damage is done. To then focus on trying to either outlaw them, or shame people out of using them, is too little, too late. Enumerating the damage, being vocal about how to fix it, and make the models useful, ethical, and fact-checking wild claims about their abilities and future to get people to start thinking about it critically, as just another product, should be the focus.

It sounds like a losing proposition on the surface, but under the deafening hype, you’ll find reality colliding with promises AI startups couldn’t keep. Company after company jumped on the LLM train and laid off people who understood the jobs being given to a test prompt, and now said companies are looking for someone to fix the messes they were warned would be the end result.

Just as so many traditional automation projects tend to fail thanks to hubris from the company’s “idea guy,” scope creep, ignoring subject matter experts, and buying into the hype of a trendy technology while ignoring its gotchas, so do AI-driven ones.

It needs to sink it for all the business owners and investors who hear nothing but how downright miraculous these generative bots are, and how they’re going to automate their companies into 99% profit margins if they just buy access to them for a measly $10,000 a month, that they’re being sold a very generic tool. That the technology in the cores of these models is useful, but that it needs to be harnessed and directed by people who know what they’re doing, how to properly train and evaluate their output, and how to fix their hallucinations and mistakes now, not with the next release.

Otherwise, what they’ll hear is that the AI is amazing and can do anything, and all the angry Luddites worried about their jobs are spreading FUD about their newest path to getting rich — or richer — quickly and with minimal effort. And as we all know, when a certain group of people is blinded by greed, any argument about ethics, morality, and empathy will fall on deaf ears.