“AI will make everyone dumber than they already are. Do you really need to ask AI what to do with your leftover ingredients?”
I’ve heard variations of this from friends lately. The worry is that we’ll get lazy and over-reliant on quick answers and stop being curious altogether. Is any of that true? If we can just “ask Chat”, do we stop asking each other? Do Google’s “AI Overviews” mean we stop questioning, experimenting, or figuring things out for ourselves? My guess is that it’s less about the tool than about us.
Whenever I hear about tech anxiety, I think of the Aeon essay “We’ve always been distracted”, which traces panic about new information tools rotting our brains back to the invention of writing itself. Plato implied in a story that writing would breed forgetfulness. In the 16th century, scholars worried that book indexes would make readers lazy because they’d skip reading the full text: the Renaissance equivalent of only reading the bold text of an AI Overview. Yet indexes didn’t ruin reading. Maybe some of us would even say that it’s absurd to think they could have.
I wonder if AI is just another tool that reflects the habits we already bring to it, particularly our curiosity. A scholar in the 16th century was presumably pretty good at reading, and maybe using an index made them more efficient, not lazier. Maybe AI doesn’t create passive behaviour. Someone who wants a quick dinner idea and asks an LLM what to do with leftover chicken isn’t losing their curiosity. They’re just taking a new version of the old shortcut: the ready-made recipe in a magazine, or the first link on Google (plus all those ads to scroll through).
At the same time, an LLM doesn’t make the decision for you any more than the magazine did; it’s still yours to accept, question, or ignore. Take choosing an exercise routine. Decades ago, people put together routines by flipping through books and workout columns. Later, we cross-referenced articles and videos on Google. Today, I might ask an LLM to generate a baseline routine around specific goals and the kinds of workouts I like. With my experience and interest in the subject, I still swap out exercises, shorten the workout, adjust the sets and reps, and fine-tune the whole thing. For anyone interested in the details of their training, it isn’t all that different from past decades. The medium changed, but my agency (and enjoyment) didn’t.
But what about someone who wants the “best full body workout”, types exactly that into Google or an AI prompt, and takes whatever comes back? LLMs are often confidently wrong, and a beginner might assume that answer really is the best routine for their goals even when it isn’t. Is that person going to “get dumber”? To be fair, AI is meaningfully different from older tools. A magazine article was obviously generic, and a search result at least pointed you to a source you could check. An LLM writes a complete, convincing answer that sounds like it was made just for you, which makes it much easier to accept bad advice without noticing.
Still, whether it’s AI-generated or a GQ article on the “Best Workout to Lose Fat”, beginners in any field have always started with generic advice. The travel brochure is now an AI itinerary. What usually gets people past generic advice is experience and reflection: figuring out why the workout routine didn’t work, why the dish tastes under-seasoned, or why the vacation was filled with tourist traps.
If someone who wants a quick dinner idea were interested in improving their cooking, I think they’d still get there. They might even do so by using an LLM and talking through the ingredients they have. AI might change how we start in the kitchen, but it doesn’t decide where we end up. That part has always been up to us, and to how curious we are.