why your boss is obsessed with a.i. past the point of sanity
Not only is the C-suite not immune to AI psychosis, they seem to be primed to suffer the worst of it as their employees duck and cover.
Now, we've talked about AI mandates at work, and AI psychosis, as well as how some people simply hand over their agency to chatbots, but what we haven't talked about is the intersection of all three. Believe it or not, this is one of the big new topics in the tech world today. Too many CEOs and executives are just way too obsessed with LLMs, even for ardent AI everywhere advocates' tastes.
Yes, sure, a lot of top corporate execs are mandating LLM usage and investing in every chatbot model thanks to relentless pressure from investors. But these are not the cases in question. No, we're talking about CEOs who found a god in the machine and now see the right series of prompts and intricate webs of APIs as the way and the light not just for their business, but all businesses in genuinely problematic ways, and are refusing to listen to feedback.
One of the biggest and most immediate problems is something being called work slop. Employees forced to use LLMs for every task until the costs of using millions of tokens a day exploded because chatbot makers are trying to go public and recoup some costs, have been really just phoning it in with AI barf that at first looks like completed work, but is actually just the first thing the agent they were using generated without so much as a basic fact check, leaving their coworkers to fix all the errors.
It's already problematic when a frontline worker does this and you now have an angry or confused customer on the other line, or lose a sale because clients see slop instead of a tailored sales pitch – while more than likely doing the same thing, but the corporate world doesn't let hypocrisy stand in the way of the illusion of cost savings – and decide to go elsewhere.
When a CEO does it across the entire business, with no regard for the last mile of work, it can be disastrous, so much so, some AI slop startups are getting worried about how their tools are being used and warning their users how their LLMs are great and so are their agents, but before implementing any suggestions, you need a sanity check and a thorough review of the outputs. Sure, this seems obvious and you'll hear executives repeat this sentiment or echo it on cue, then just... not do it.
Of course, AI companies aren't worried because they have our best interests in mind, but as a PR move as more and more employees are watching their bosses hand over their jobs to ChatGPT or Claude, relying on these chatbots to make decisions that used to be made by teams of subject matter experts, then issuing conflicting and whiplash-inducing edicts on a weekly basis to their ever more confused staff.
But there do seem to be three reasons why CEOs and executives are obsessed with AI, and why it's going to be very difficult to treat their addictions because these factors appeal to the key aspects of an exec's mentality, seemingly tailor made for nudging them deeper and deeper into AI psychosis.
First, it appeals to their desire to be praised and agreed with, and to be treated like the best and smartest person in the room, and chatbots are trained to love bomb you into near oblivion no matter how wrong you might be, and can very easily be convinced to validate your worst ideas. For people who rarely, if ever, hear the word no, this is very appealing because look, a machine made of nothing but math and logic thinks they're just so smart and have such amazing ideas.
Second, one of the biggest challenges of higher level work is the lack of instant impact. You are now removed from day to day operations and the results of your work are felt over weeks, if not months or years. You suddenly need buy in and consensus, and your policies and methods need to permeate and go through a few iterations before you see payoff, that also arrives at the speed of a narcoleptic snail.
But the latest suite of AI tools allow you to create a virtual shadow company staffed by machines that lavish you with phrase and simulate your big ideas as if they're real world pilot projects. It looks and feels like you're getting an immense amount of work done at tremendous speed, even if in reality, you're basically just playing a very expensive text version of a business tycoon simulator. Think of it as the boss experience gamified, and with near-instant gratification.
Third, is that it gives them a tool to help generate new ideas, but also blame when they don't work out. Consider that Facebook never really evolved past being a minimalistic MySpace with better targeted ads, and any attempt to diversify past that involved either buying another platform or projects no one asked for or wanted to be a part of, like AI friends, the Metaverse, impersonating the dead with AI, and glasses that record your most intimate moments to be watched by underpaid contractors to train AI models.
All horrible, terrible, tone deaf ideas that could not read the room worse because in mid-research and development, the room kept loudly screaming that they hated all of this and just wanted Meta to stop polluting their feeds with rage bait. So, if Zuck and the other Meta big shots delegate business proposals to an LLM, it may just come up with something that works. And if it doesn't? It's the bot's fault and they just need to train the magic machine to figure out how to make more money. That's the whole pitch for modern LLMs after all.
We could unpack what all this says about powerful people, their shared quirks, and the incentives of the jobs they do, but I think the easiest way to sum it all up is that AI lets them feel productive and accomplished by feeding them the illusion of both, all while stroking their egos. That means infatuation with LLMs is going to be very hard to break and may not happen at all until using them too much becomes a very clear and obvious liability enough times that "but this time, in my hands, it's going to be different" is an excuse nether customers not shareholders will accept.
Even worse, this means that the whiplash you keep getting from your boss' wild pivots after a long night with their favorite chatbot is here to stay for the time being, and odds are that it will get worse before it gets better.
