In Full...
For as long as any of us have been alive, you trusted the expert. You went to the doctor, or the solicitor, and they told you what was what, and you took their word for it. You didn’t really have a choice. You didn’t know enough to argue, you’d no way of checking, and questioning the professional wasn’t the done thing anyway. So you nodded, and got on with it. That deference held a surprising amount together: a kind of trust we’ve handed to institutions for decades without thinking about it much.
That’s starting to come apart, and not only because of AI. Faith in the experts has been wobbling for years - Covid did a real number on it, with the whole country watching them contradict each other and change their advice in real time. Although that cut both ways. Plenty of the people who decided to ignore the experts altogether didn’t exactly cover themselves in glory. Sometimes the expert really is right, and the confident bloke online really isn’t.
AI does something different, though. It lets you actually check. You come out of the appointment, go home, and ask it what it reckons about what you were just told. Loads of people are already doing exactly this. It feels like a small thing, but it isn’t: you’re second-guessing the professional in a way that simply wasn’t available a few years ago, for free, on your phone, with nobody to feel awkward in front of. And some of the time it’s genuinely useful. Experts get things wrong, the symptom waved away or the advice that was a bit off, and now you can at least ask the question you’d never have known to ask.
It isn’t all good, though. There’s something a little corrosive about a world where nobody quite trusts the professional any more, where the doctor who’s right ninety-nine times gets fact-checked on the hundredth. And the thing we’re checking with isn’t exactly trustworthy itself; it’ll give you something completely wrong in the same confident voice it uses for everything.
The AI itself is built partly out of doctors, and not average ones. The companies behind these things pay top people in each field, leading doctors and lawyers and the rest, to sit and write the answers that train the models. So when you check your GP against it, you’re in a way reaching past them to the best in the field. That’s not nothing. But you get it all as one flat, confident summary, with the bit that tells you how much to trust any given part of it taken out.
I don’t think there’s a clean answer here. It’s a good thing and a bad thing at once. We spent a very long time trusting the people in the room because we had no real alternative. Now we have one, of sorts. I just can’t tell yet whether it leaves us better off, or simply more alone with decisions we were never that well equipped to make in the first place.