alt Hacker NewsWhatever the tool, the problem is that it’s wrong, not that it’s done by AI. Like, why are we conflating the two - is it because it’s now easier for someone to produce a wrong artifact that visually looks ok and therefore tricks our brains? And people get upset because a prior quality heuristic is now obviously broken?
Replies
Its because AI makes it 100x easier for people to make things that are wrong, and feel like they got it right. It is a dunning-kreuger amplifier.
Yeah I think so. I mean it's a new social norm - previously if you replied with "hey I read this article here and it says X" it would be really easy to check the source and either say yeah this looks good or disprove it.
Now people will say "I asked AI to give me a map of ocean currents and it definitively says the person is here!" and the argument that "look it's AI, it's probably wrong" just does nothing for a lot of people. Doubly so if the information they want is at the forefront of Google, after all, Google is a big reputable company, so of course what it says is correct.
Yes I agree that the AI itself isn't the problem, but the way it's being pushed and promoted(just ask anything!) is leading to this.
Because this new tool will also tell you, in (mostly) human language, why the wrong artifacts it createds are correct. And it's arguments are very convincing to a lot of people. AI is clearly creating new problems that didn't really exist before, you can't say it's just a tool like any other when this is the only tool where its users form parasocial relationships with it.