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

131–140 of 486 posts

Re: Conservatives think ChatGPT has gone 'woke'

#131

Post by fidgewidge »

loudmax wrote: post #77
RC_ITR wrote: post #25

Earlier quoted context omitted.

Divorcing the conservative/liberal split from its current muddied use in American politics: Conservatism generally follows the principle of “be conservative in your attempts to alter society”. OpenAI is being aggressive in moderating ChatGPT, and that’s against the core principle of conservatism (at the end of the day, LLMs are taking what people say and reflecting it back, but OpenAI is adding the extra step of only…

The meanings of "Liberal" and "Conservative" with respect to American politics are completely haywire. At the extreme ends we have a far left pushing illiberal restrictions on free speech, and a far right cult of personality inciting mob violence. Referring to those extremes a liberal or conservative is misleading. That's not what those words mean.

Trump isn't particularly right wing let alone far right. This is probably still the best takedown of that idea:

https://slatestarcodex.com/2016/11/16/you-are-still-crying-w...

Trump is politically/ideologically center left. He has very little to say about wokeism, was fine with vaccine mandates, and said things like this:

"America must reject the bigotry of Hillary Clinton who sees communities of color only as votes, not as human beings worthy of a better future."

Also he was a Democrat in the past.

The term far right doesn't make any sense if you think about it for a second. It's not just in American politics. People describe the NSDAP as "far right" even though it was largely indistinguishable from the USSR which everyone agrees was far left. Far right would logically be the extreme inverse of communist countries like China or the USSR: shrink the government at any cost, freedom of speech without limits, repealing laws en masse, refusing to take over the world and so on. So extreme libertarianism. In practice though, this isn't what people mean when they say far right.

Re: Conservatives think ChatGPT has gone 'woke'

#132

Post by wizeman »

karpierz wrote: post #126
wizeman wrote: post #115

Earlier quoted context omitted.

> it's always going to be biased towards some direction, whether that's the views of the society it pulls most of its data > I don't think there's such thing as a lack of bias If the AI is simply reflecting the data it was trained on and this data is a representative sample of all data, isn't it unbiased by definition? I don't think we should just throw our hands up and say "this is impossible" just yet. That's just…

> If the AI is simply reflecting the data it was trained on and this data is a representative sample of all data, isn't it unbiased by definition? It's unbiased by definition of "does the output reflect the input"? It's not unbiased by definition of "does the output reflect reality"?

> It's not unbiased by definition of "does the output reflect reality"?

How does "all data" differ from reality?

Re: Conservatives think ChatGPT has gone 'woke'

#133

Post by breadbreadbread »

unbiased AI is literally impossible. The nature of data labelling, and even language itself, means that somewhere a human is deciding what an AI "sees". If you want to make a transphobic AI, you can do that, no one is stopping you. You just have to label data objects according to your shitty worldview. Boohoo, someone decided not to let their creation be used to spread election misinformation or scaremonger about trans people.

Re: Conservatives think ChatGPT has gone 'woke'

#134

Post by fidgewidge »

everdrive wrote: post #102
vkou wrote: post #62

Earlier quoted context omitted.

I don't see why it's an AI issue only. Imagine how awful it would be if millions of people got their information from a biased carbon-based neural network, like Tucker Carlson [1]... Is there something that we should do to prevent such a problematic outcome? Is it really a good idea that clearly biased information is being broadcast to millions of people? [1] The entity that appears on television known as Tucker Carl…

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.

Re: Conservatives think ChatGPT has gone 'woke'

#135

Post by scarmig »

wizeman wrote: post #130
scarmig wrote: post #99

Earlier quoted context omitted.

If I were a corporation looking for an LLM for some product feature, I would absolutely go for the one with more "woke" opinions, even if if resulted in a worse customer experience. If you didn't, you risk a lot of media and government backlash. It's all about context.

> If I were a corporation looking for an LLM for some product feature, I would absolutely go for the one with more "woke" opinions, even if if resulted in a worse customer experience. How about instead of preferring the LLM with "woke" opinions, you would prefer an LLM that was simply trained to avoid controversial topics? That way, you could use it for your product while still avoiding both bias and media/government…

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 about free speech and open dialogue.

Re: Conservatives think ChatGPT has gone 'woke'

#136

Post by pixl97 »

MuffinFlavored wrote: post #35
snicker7 wrote: post #28

As described in the article, the political bias is intentional. It is the result of (not-so-transparent) ethical guard rails baked into the system.

Is that to say at a higher level "liberal bias = ethical/ok for boat guard rail training material, conservative bias = typically unethical and to be avoided"? I feel like that plays into conservative hands of "they're trying to silence us!" Why do 70,000,000 people vote (R) every year, knowing that the other 70,000,000 (D) think they are "unethical"?

Because the 70M(R) want to enact their own set of 'unethical' laws that would greatly affect the (D).

Unless you can think of an actual ethical reason that gay marriage should be banned?

Re: Conservatives think ChatGPT has gone 'woke'

#137

Post by pohl »

dunste wrote: post #129
pohl wrote: post #118

Earlier quoted context omitted.

those overrides tend toward privileging fashionable American left-wing positions ...such as whether trans people are deserving of equal rights, or whether or not the 2020 election results were fraudulent (which, if one reads TFA, were the cited complaints)

[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).

Re: Conservatives think ChatGPT has gone 'woke'

#138

Post by dunste »

breadbreadbread wrote: post #133

unbiased AI is literally impossible. The nature of data labelling, and even language itself, means that somewhere a human is deciding what an AI "sees". If you want to make a transphobic AI, you can do that, no one is stopping you. You just have to label data objects according to your shitty worldview. Boohoo, someone decided not to let their creation be used to spread election misinformation or scaremonger about tra…

ChatGPT already gives an answer that would be considered 'transphobic' by those who hold strong ideological beliefs on the primacy of gender identity over sex:

> What is a woman?

> A woman is an adult female human. Women are typically characterized by their reproductive biology, which includes their menstrual cycle, and the production of eggs and estrogen. Women are also known for their unique physical characteristics, such as their breast tissue and hips. They are also often distinguished by their social and cultural roles, such as their expected behavior and appearance. In general, the term "woman" is used to refer to any adult female person, regardless of her age, race, ethnicity, or other characteristics.

Re: Conservatives think ChatGPT has gone 'woke'

#139

Post by wizeman »

scarmig wrote: post #135
wizeman wrote: post #130

Earlier quoted context omitted.

> If I were a corporation looking for an LLM for some product feature, I would absolutely go for the one with more "woke" opinions, even if if resulted in a worse customer experience. How about instead of preferring the LLM with "woke" opinions, you would prefer an LLM that was simply trained to avoid controversial topics? That way, you could use it for your product while still avoiding both bias and media/government…

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.

Re: Conservatives think ChatGPT has gone 'woke'

#140

Post by scarmig »

wizeman wrote: post #132
karpierz wrote: post #126

Earlier quoted context omitted.

> If the AI is simply reflecting the data it was trained on and this data is a representative sample of all data, isn't it unbiased by definition? It's unbiased by definition of "does the output reflect the input"? It's not unbiased by definition of "does the output reflect reality"?

> 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.
Post reply on HN