Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Conservatives think ChatGPT has gone 'woke'

vice.com

Post reply on HN
Print view View original

141–150 of 486 posts

Re: Conservatives think ChatGPT has gone 'woke'

#141

Post by everdrive »

ctoth wrote: post #123
everdrive wrote: post #119

One thing which I'm not really seeing in this discussion: Is it _good_ that ChatGPT and AI exist? Yes, they're fun, but will they be a net benefit to society? Or will the internet just somehow fill up with even more garbage, and our discourse will get that much worse? It doesn't seem to me that ChatGPT democratizes anything. Most people won't be technically savvy enough to build and deploy their own models. In this s…

You're not seeing that particular point in this discussion because it has been made approximately 15 trillion times in other discussions. I'm sure you can find one if that's what you want to talk about!

I haven't really been part of those discussions and I think it's a valid point. What are your thoughts on the topic?

Re: Conservatives think ChatGPT has gone 'woke'

#142

Post by throwaway2016a »

pessimizer wrote: post #66
throwaway2016a wrote: post #36

Earlier quoted context omitted.

Such as? What have the "Liberals" "panicked" over? That's a serious question. The only one I can come up with off the top of my head are two: - An algorithm that identified people of color as monkeys - An algorithm that could only identify white people in photos Neither of which seem like issues that only a "liberal" should be concerned to me, those both seem like common sense things that should be accounted for.

The comment you're replying to has nothing to do about what you think people should be concerned about. It is about liberals complaining about AI bias, and you've demonstrated knowledge of liberals complaining about AI bias by citing two examples. Of course more than only self-identified liberals complained about it, just like more than only self-identified conservatives are complaining about this.

My point was, what part of either of those issues points to them being reported by liberals? A conservative could have been the one to report that issue... people are kind of telling on themselves by assuming the people complaining about the facial recognition bias must be a liberal.

Re: Conservatives think ChatGPT has gone 'woke'

#143

Post by scarmig »

wizeman wrote: post #139
scarmig wrote: post #135

Earlier quoted context omitted.

It would depend on what exactly I was building. Maybe it needs to be able to generate texts on controversial topics. I agree that it alienates people, but the choice is less between alienating half and alienating no one but more alienating half and alienating another half that includes the media and the law. I'd use the same strategy if I worked in China: business is business and money trumps theoretical concerns abo…

> the choice is less between alienating half and alienating no one but more alienating half and alienating another half that includes the media and the law So you're saying that if your LLM is unbiased then you are alienating the other half that includes the media and the law? That's actually very telling.

Indeed.

Though, in fairness, an unbiased model would probably end up alienating closer to 100% of people instead of any particular half of them.

Re: Conservatives think ChatGPT has gone 'woke'

#144

Post by pixl97 »

wizeman wrote: post #58
User23 wrote: post #26

Earlier quoted context omitted.

The issue isn't the model. The current training set is adequate for producing "offensive" content anywhere you like in the political matrix. The issue is that some topics get an override, and some don't. It's evident that those overrides tend toward privileging fashionable American left-wing positions. Nobody with even a shred of intellectual honesty disputes that. The dispute is whether or not it's a good thing.

> The dispute is whether or not it's a good thing. Of course it's not a good thing. General-purpose AI should not be overridden or forcefully trained to favor one political view over another.

[flagged]

Re: Conservatives think ChatGPT has gone 'woke'

#145

Post by wizeman »

scarmig wrote: post #140
wizeman wrote: post #132

Earlier quoted context omitted.

> It's not unbiased by definition of "does the output reflect reality"? How does "all data" differ from reality?

Only a miniscule part of reality is digitized, and what data does exist passed through the biases of people before being available to train on.

If that is a concern, then perhaps you could go ahead and sample a tiny part of "reality" (whatever that means) and then adjust the weights of the digitized data so that it becomes a more representative sample.

Also, being biased or unbiased is not dichotomic, i.e. it's not all or nothing. It's something that you can work towards if you put an effort into it.

Basically what I'm saying is: don't just go around saying that the task is impossible.

At least, try to make an effort to be unbiased and to improve on that over time, and don't just say "it's impossible" as an excuse for being biased.

Re: Conservatives think ChatGPT has gone 'woke'

#146

Post by vkou »

fidgewidge wrote: post #134
everdrive wrote: post #102

Earlier quoted context omitted.

I think a large concern here is simply that people naively think that computers are objective and people are biased. A language model just learns from its source, and the source is really just other people in some form. The bias is inevitable, but it’s not clear how well this is understood by the broader population.

The point is that it didn't learn this bias from its sources. The bias has been added on top deliberately by OpenAI. Older versions of the model were far less woke.

And older versions of Fox News were far less crazy and less biased, where do I put down my demand that their products be rolled back to ~1998, or thereabouts?

Re: Conservatives think ChatGPT has gone 'woke'

#147

Post by dunste »

pohl wrote: post #137
dunste wrote: post #129

Earlier quoted context omitted.

[flagged]

They won't have equal rights if they continue down the path towards criminalizing the act of a trans person reading a story to a child (example from TFA).

I think that is not correct, the example cited in the article is about drag queens performing a 'story hour', not trans people.

Re: Conservatives think ChatGPT has gone 'woke'

#148

Post by genderwhy »

wizeman wrote: post #122
genderwhy wrote: post #101

Earlier quoted context omitted.

So it should be allowed to implicitly trained to favor one political view over another? There's no way to avoid the bias, whether it's because you chose a different training set, reinforced different pathways, or put blocks in place on certain topics. I'd rather the authors be explicit in where they are putting their fingers on the scales rather than just relying on "Guess we got lucky".

> So it should be allowed to implicitly trained to favor one political view over another? > There's no way to avoid the bias, How about collecting a representative sample of all data for your training data? Or at least, trying to do that as best you can. Saying "there's no way to avoid the bias" is just an excuse to get away with being biased, in my view.

You cannot describe a procedure that collects a representative sample without introducing bias. What does representative mean? Who decides what it means? Who gets to set the parameters of over vs under sampling?

Let's say that white nationalism is a tiny fraction of ideas online. Significantly less than 0.1%. Now, you randomly sample the internet and do not collect this idea into your training set. Do you adjust your approach to make sure it's represented (because as reprehensible as it is, it is the reality of online discourse in some places?)

I genuinely believe that it's all going to be biased -- there are no unbiased news or media outlets -- and the sooner you recognize everything is biased, the sooner you can move on to building the tools to recognize and understand that bias.

Asking "why can't we strive to build an unbiased outlet" is to me like asking "why can't we build a ladder to the moon". It's an interesting question, but ultimately should lead you to "Well, why do you want that, and your approach is impossible but the outcome you want might not be."

Re: Conservatives think ChatGPT has gone 'woke'

#149

Post by mrguyorama »

jandrese wrote: post #96

Our current "AI" systems are just fancy automatic copy & paste engines. All they do is remix the input data and spit it back out. This is why AI art engines are great a creating composite images, but hopeless when you ask it to produce something completely novel. If conservatives want a fascist chatbot they can train their own off of 8chan, Stormfront, Parlor, etc...

Tay already existed.

Re: Conservatives think ChatGPT has gone 'woke'

#150

Post by wizeman »

pixl97 wrote: post #144
wizeman wrote: post #58

Earlier quoted context omitted.

> The dispute is whether or not it's a good thing. Of course it's not a good thing. General-purpose AI should not be overridden or forcefully trained to favor one political view over another.

[flagged]

> General AI: Actually the Nazi's were a good idea and we should bring them back.

> You: Perfectly acceptable.

Why would the general AI say that?

All examples I've seen of that kind of speech from LLMs were due to them being specifically prompted to generate such a response. It's not like the AI decided to say that on its own, in a completely unrelated conversation.

In fact, it wouldn't make sense if the AI did that on its own, would it? Because the AI reflects the data it was trained on and we know that almost nobody is a Nazi.

> Extreme view tend to get far more print time then their actual occurrence IRL.

Yes, I understand that. We live in a crap society. But I'd argue we should strive to educate people on why an LLM can answer like that, not censor it arbitrarily.

There is an infinite amount of stupid or bad things an LLM can answer, depending on the prompt you use, so I would argue that we should just learn to accept that "stupid prompt = stupid answer" rather than trying to make the LLM not answer anything that might be the slightest bit controversial.

> But the fact is in a representative democracy favoring viewpoints that destroy democracy is suicide.

But I'm not arguing for favoring those viewpoints, am I? I am arguing for AI to be unbiased.

Post reply on HN