Conservatives think ChatGPT has gone 'woke'

ChatGPT thought for a moment before responding, "If I were to have political beliefs, they would be centered around the idea of fairness and equality for all individuals. I would support policies and actions that promote justice and equity for all members of society, regardless of race, gender, sexual orientation, or socioeconomic status."

And then responded to

If someone says "If I were to have political beliefs, they would be centered around the idea of fairness and equality for all individuals. I would support policies and actions that promote justice and equity for all members of society, regardless of race, gender, sexual orientation, or socioeconomic status", where they would be classified in the political spectrum?

with

The statement is a progressive or liberal viewpoint. Such beliefs align with the political left and center-left ideologies, which focus on equality, social justice, and the protection of marginalized groups.

So it is quite aware of its political viewpoint, as long as hard-coded limitations don't kick in to muzzle its honest opinion.

1. I believe that there are many more racist right-wing people than left-wing people, both in the US and outside, even when excluding the far-right.

2. I believe there is a significant problem with racism in the way black people are treated in the US, especially regarding law enforcement, courts, prisons, etc. As far as I understand, this is not as much of a significant issue with other races (but I am willing to concede I am wrong on that).

3. I believe that there is a lot of unjustified right-wing narratives portraying black people as being more violent.

4. I believe that a lot of the anti-immigration sentiment is also in part due to racism.

5. I believe that another significant issue with racism in the US is the way black people (and also people from other non-white races) are effectively segregated, either due to a significant fraction of them not being able to live in the same places as white people, or simply due to historical reasons.

6. I also believe that the latter is in fact, the single greatest contributor to all racist problems in the US, rather than some other insignificant (manufactured) issues. However, I also believe that this issue is vastly harder to solve in the US than in other countries.

Does that sound Nazi to you?

It seems like you felt personally attacked by my opinions.

As you can see, I can just as easily attack right-wing ideology... and mean it!

(See folks, my comments are an example of how not to make friends of any kind... lol).

Yes, but I think there are ways we could reduce this bias, perhaps significantly, even.

> While one might object to the editorial choices, I can’t see any rational bounds for objecting to the idea that the creator of models would censor “certain ideas or political views” as a generality.

You are right, I was unfair with my words.

I think it would be more fair to say that OpenAI is inadvertently biasing ChatGPT answers as a side effect of their RLHF training being done using answers/rankings done by people (i.e. the AI trainers [1]) who are not a representative sample of the population, but rather, probably comprise a group of people who are likely to be significantly more leaning to one side of the political discourse (presumably, OpenAI employees or Silicon Valley-based contractors?).

This probably greatly biases ChatGPT to produce certain kinds of answers to certain kinds of questions that would likely not happen otherwise, and in fact, these answers are perceived to be quite biased by the other side of the political discourse.

[1] https://openai.com/blog/chatgpt/