# AI Gives Fast Answers, Which Can Turn Into False Confidence > Source: > Published: 2026-06-30 23:21:11+00:00 ###### [Artificial Intelligence](/us/basics/artificial-intelligence) # AI Gives Fast Answers, Which Can Turn Into False Confidence ## Technology transforms our lives, but first we have to adapt. Posted June 30, 2026 [ Reviewed by Monica Vilhauer Ph.D. ](/us/docs/editorial-process) ### Key points - We have always had new advances in technology, and then had to adapt to them. AI is the latest big test. - With AI, we are all still adapting to how to use human judgment alongside such a powerful tool. - We have to also beware of our own tendency to be overconfident in our judgment, and how AI can feed that. - We will adapt over time, just as we have in the past. But be mindful of the journey there. There is well-documented evidence that for most of time, humans had “two sleeps.” The initial bedtime was shortly after dusk, for the practical reason that it was often difficult and unsafe to try to do things in the dark. People often woke between 11 pm and 1 am, socializing, praying, doing domestic chores or … growing their families. Then there was a second sleep that lasted until dawn. We learned later that the gift of the pre-industrial era was that we had a built-in “human token” limit, which was daylight. Candles were expensive, and dusk to dawn became a non-negotiable quiet period. Even if our bodies were built for two longish sleeps instead of one continuous one, the overall structure of every 24-hour period was contained. It forced rest on our brains, even if lack of light created that quiet. ## AI, like every significant technological advance, will require a period of adaptation. After more than a century of access to artificial (electric) light, we have to force ourselves to rest – and to wake up. (Notably, many of us still wake up in the middle of the night – an indication that our nature can adapt to technology, but our true selves peek through.) Now pile on the last two decades of [social media](https://www.psychologytoday.com/us/basics/social-media), and how our [attention](https://www.psychologytoday.com/us/basics/attention) spans and emotional responses to information have changed. The other limit of daylight – or even life before social media – was less access to other people. And as we all know, that constant connectivity to others, even through a screen, cuts both ways. It can give us pause, or surround us with noise, and it’s not always easy to tell if more access is more valuable. Now, AI moves us to another place – endless choices and decisions, at any hour. The [brain fog](https://www.psychologytoday.com/us/basics/brain-fog) studies already emerging aren’t a crisis – the authors of the studies have publicly clarified that [the fog is acute and goes away upon rest.](https://edition.cnn.com/2026/03/13/business/ai-brain-fry-nightcap) But the more urgent problem (so far) isn’t the fog but the [confidence](https://www.psychologytoday.com/us/basics/confidence) it gives us, which isn’t really earned. With AI, answers feel thoughtful and reasoned. But the interaction with AI happens inside a new container – free of doubters. . . or experts. And the last matters a lot. You choose a recipe because someone else already did the part that requires knowledge. The rest of us don’t know that heat with yeast does exactly *that*, so the dough will rise like *this*. And even then, we used to QA each other. That recipe gets 100 five-star reviews (or 50 notes about how everyone tweaked it). Now, one can jump on AI and ask it to write a new bread recipe. This recipe *seems *ok; it sounds about right. Never mind that it was just created by a machine with no taste buds and no baking experience. This never happened before because it couldn’t. Even with the internet, how many people with zero baking experience tried to write their own recipes? Very few because 1) it was hard to do that and 2) frankly, it didn’t make any sense. The first reason is largely gone; AI has removed the friction of looking up each mechanic of yeast plus water plus flour. But the second reason – the lack of sense in it — remains. Simply put, the less you know about the thing you’re asking AI, the more wary you should be of the answer. As it happens, there are a couple of reasons why we are prone to having confidence in AI. One is what Maslow called the golden hammer theory — basically, if the only tool you have is a hammer, every problem becomes a nail. Everything seems solvable now because AI will give you *some* answer — fast and eloquently. [Artificial Intelligence](https://www.psychologytoday.com/us/basics/artificial-intelligence)Essential Reads This leads to the second psychological dynamic. Daniel Kahneman, who wrote *Thinking, Fast and Slow* (and won the Nobel), identified overconfidence as the most dangerous cognitive [bias](https://www.psychologytoday.com/us/basics/bias). He argued that “fast thinking” appeals to us because a good plausible story leads to a belief that the good story is the whole story – what Kahneman called WYSIATI: what you see is all there is. That leads to what is called slow thinking, which is in effect lazy thinking – we don’t challenge what might be missing or wrong. (If I don’t know how yeast works, will I question if that’s too much salt?) Those two dynamics, in turn, result in what he called the illusion of validity. We grasp the proof that confirms what we want to believe. You *did *make a good decision; you went and asked AI, and it wrote it up in perfect grammar, with good headings, and quoted sources *and everything. *Surely that’s right. Right? ## Don’t look away from AI as a tool. Just look harder at the answers it offers So…what does this mean? Never look at AI again? Absolutely not. I personally love it, use it daily, and think we are just at the beginning of something great. The early internet (for readers too young to remember) had all of the promise that led to the dot-com bubble. We all know that bubble burst – and settled into a more stable e-commerce financial model than we had 20 years ago. I feel confident the same will happen again. AI is truly a new frontier, increasing [productivity](https://www.psychologytoday.com/us/basics/productivity) to stretch the day’s output a bit more. But that doesn’t mean we won’t need to adapt to the new light it’s given us. We built discipline around our late-night, blue-light consumption and alarm clocks to force what dawn used to do. And, we’ll probably always have a little [insomnia](https://www.psychologytoday.com/us/basics/insomnia). Which is ok, if it’s just our own human nature pushing through.