I’ve been complaining about AI slop, probably not publicly, but clearly. And, I’ve been noticing the complaints about brain rot from using AI. Let’s be clear, this is an instance of people saying “AI” and meaning generative artificial intelligence, a subset. Still, both are being seen, and not surprisingly, they’re linked. At least, let me make the case that AI slop and brain rot are related.
AI slop, to be clear, is when folks are using AI to generate content. It can be passive, as that awful image that had me as one of the ‘wealthy influential L&D leaders’. Wealthy? As if (as in, in my dreams). Sorry, my bad habit of ‘standing up‘ for what I believe gets in the way, way too much! AI slop can also be interactive, such as the chatbots (textual or auditory) that are replacing people. In either case, it’s using an average of stolen intelligence to create responses, and that’s only minimally worthwhile even if, as Markus Bernhardt reminds us, we provide oversight. Which doesn’t happen, because it’s too cheap to just have AI create worthless content.
There is also the phenomena of brain rot, where people who use AI, unless using it well (which is rare) get dumber. If you outsource the mental challenge, you aren’t doing the work that maintains and builds mental capacity. We’re increasingly seeing evidence that folks who outsource to AI end up diminishing their capability. Unless they’re using it as an idea partner and a form of feedback – generating their own ideas and using the AI output as a check – they’re setting themselves up for failure. It’s like with your muscles, use ’em or lose ’em. Same with your brain.
Similarly, companies that let AI handle low-level tasks are breaking up the pathway to expertise. Sure, it’s a short-term win, but…what happens when the expertise retires, and you’ve no one who’s been developing to replace? Maybe AI will get good enough to handle that. But there’s also evidence that the technology is reaching the law of diminishing returns. And, what with the move to the market, the costs are likely to go up. Finally, the tech is optimal for language tasks; using it to do other things, such as numeric tasks, is a quick trip to inaccuracies. There are better approaches.
At core, my point is that the source of the problem is the same: the technology isn’t being used in ways consonant with it’s strengths. (Let alone in ways consonant with society as a whole!) We’re not using it to augment us in ways that are beneficial, instead we’re using it in ways that produce short-term savings at the expense of long-term benefits. And that, to me, is a path to a worse future, and I think we can do better.
Overall, the rush to ‘AI’ is pretty short-sighted. Like most things, we see an overexcitement, and then we face the reckoning. The problem is that this rush has been substantially bigger than any ever before, and the consequences of the resulting rationalization could be onerous. Maybe it’s time, as we’ve seen before, that we slow down and look at how to use this new tech appropriately. I like a meme I just heard: instead of move fast and break things, move slow and fix things. Is it time?