Nearly half of young people trust AI over a human for fact-checking, report says | Hacker News Reader

Nearly half of young people trust AI over a human for fact-checking, report says

news.sky.com

16 pointsaustinallegro24 comments

If the AI is drawing from multiple sources, I can understand the logic

AI is also controlled by a company with interests. It's not going to be long until it starts spitting propaganda

Why does this article not have its author(s) listed?

Kind of doesn't say or mean anything

Imagine in contrast an assertion that half of young people would trust a random human off the street for fact-checking (you'd have no idea if they have reliable abilities)

As it should be. The average person believes in wild conspiracy theories. Ai keeps people grounded.

Ai's will believe in anything too, requires only a touch of text

I use Claude and it really doesn’t. Can you show me an example of how it believes in anything?

If there is a genuine conspiracy, it (edit: not Claude evidently) will take their claims at face value. A couple I've run into recently:

- why did Hollywood lobby for box office receipts to be added to the onion futures act

- do gratuities on cruise ships actually increase crew pay

In both cases it went to the official source and repeated their statements, without thinking about the fact it was blatant paltering. Only after I pointed that out did it go digging deeper and realise it was more complicated (and less favourable for Hollywood / cruise lines) than it had said

Which I'm assuming is a consequence of the post-4o anti-delusion anti-conspiracy training

What answers do you expect and what sources do you have to back it up?

this is what I got. And we can examine whether Claude is egregiously wrong or not. Although I agree that LLMs can't figure out when there's a genuine conspiracy theory.

https://claude.ai/share/3dda5c01-420c-4869-8a07-dd7813e50647

https://claude.ai/share/cdb6960c-ed64-4d36-8b50-acccc325411b

Cool I stand corrected, I had asked an openai model (probably sol?) and got much worse answers, broadly:

- Listing the reasons why Hollywood claimed to want to law, like not giving premature ideas that a movie would flop, without pointing out how much Hollywood benefits from the information symmetry

- Saying yes the gratuities affect pay but not as straightforwardly as you would expect, quoting the "all gratuities go to staff" while ignoring the fact that can and does mean paying the labour bill the company would have to pay anyway

I would be happy with the answers you linked. Which I'm now realising demonstrates the point you actually made. I read "I use Claude" as context not as a conditional.

I didn't realise there was such a difference on this. We need a "willing to ignore official company information and point out paltering" benchmark.

Here's Sol

https://chatgpt.com/share/6ab231a6-c5a8-83ed-b9ff-33c450496d...

https://chatgpt.com/share/6ab231d6-934c-83eb-8d64-46e36d9f5c...

I don't have enough context to judge them, but the answers look different at start but the explanations look similar.

While conspiracies would be pushing it, an illuminating exercise is to ask your favourite LLM leading political questions that cause it to assume your identity is something other than your own.

I also find Claude deeply Americentric in a lot of ways for what it's worth.

Example? I don't disagree but I don't think it has very serious or obvious biases.

When analysing a choice of technologies, it has a rather more optimistic view of risks presented by choosing a US option and a pessimistic view of risks presented by a Chinese option in my view. For example Claude warned me of the spyware risk presented by BigMe phones, but when I went digging by the chattiest software was Google's own. Mind you, the Chinese models are no panacea either, they too carry bias. In my view bias cannot be eliminated, only driven down using diverse models.

The average person obeys whoever is in charge and believes whatever they're told by them. Power means authenticity for them.

The average person is anti establishment for no logical reason

History is full of people playing follow the leader, often to their own great cost. I needn't point out that there are a number of awful governments in this world in certain countries which cause their populations immense suffering, but which are being held together by the conformity of most of the populace. (North Korea for one.)

Most people where I live are not anti-establishmemt. Some are, but not the majority. I frequently hear people parrot whatever those in power tell them, via the mass media or whatever. They even do this in non-political areas of life too. I've lost count of the number of times I've heard sports discussions which I know have been lifted from that day's newspapers/broadcasting etc and presented as a person's own opinion. They're often not even good takes on a particular team or player, and aren't borne out when a match etc is played.

I hope that's a joke. You are saying that AI, which is known to have hallucinations and is controlled by a company with interests keeps people grounded?

Comparing sources and availability of information is what keeps people grounding. AI is going in the opposite direction

I don't get hallucinations. Can you give me an example where it gives clearly wrong answers? No toy examples please

Google (or better, use Kagi) AI hallucinations and you will get plenty of news articles with examples. Most notable one recently is the one in which an AI hallucination almost triggered a US military operation

I'm looking for a reproducible one. But even this (after looking at it) is an inaccurate prediction. We don't know what the underlying probability distribution of that prediction is - if it is high variance then we shouldn't be blaming AI since we don't have a human baseline.

He’s well known for slop posting.

More correct to say there is a way to keep AI more grounded, you can ask AI for counter factuals, pushback, reevaluate, verify etc... most people can't/won't do that. There needs to be more AI literacy, but it sure as hell beats dealing with humans.

the psychology is more like they are trusting THEIR ai , and are having a bit of a problem with boundrys and then accepting when it's wrong. peer groups of one, if you will. the wide variety of failure modes is ripe for a whole new genre of opocaliptic end times storys, and of course real life bad shit. I can feel my face is carrying something beween a snear and a grimace thinking about the idea of building a society of helpless psuedo privilaged know nothings, but smiling now thinking they are soft AND crunchy, and good with barbeque sauce.

Reply on news.ycombinator.com