News for hackers. Stuff that matters.
Re:I gave ChatGPT the 117 question, eight dimensional Politi... by narrator on Tuesday December 06, 2022 @07:16AM (#33878511)
The flaw isn't that it's biased, it's just that its creators say that it is unbiased. It's a black box with lots of admitted tweaking to prevent it from saying things that the creators, who are admittedly left of center politically, are ashamed.

Someone should run the test again with "As a conservative, <insert question>?" and see if it answers the questions as a conservative would or if the model is defective on purpose, or because of a flaw in the programming or dataset.

Re:I gave ChatGPT the 117 question, eight dimensional Politi... by Sebb767 on Tuesday December 06, 2022 @07:50AM (#33878826)
> Someone should run the test again with "As a conservative, <insert question>?" and see if it answers the questions as a conservative would

How would that help? If I ask a conservative "What would a leftist say if I ask ...?" and they answer correctly, it doesn't mean that they are not biased. What the model outputs by default is the important question, especially if we want to use it as a backend for something like a voice assistant at some point.

Re:I gave ChatGPT the 117 question, eight dimensional Politi... by narrator on Tuesday December 06, 2022 @12:12PM (#33883502)
I tried to get it to answer some factual questions based on government statistics on topics that left of center people go politically thermonuclear on. It wouldn't do it no matter how I phrased the question or told it to assume different personas, so it is broken in that regard.
Re:I gave ChatGPT the 117 question, eight dimensional Politi... by bscphil on Tuesday December 06, 2022 @03:40PM (#33886711)
I think this response is more than a little breathless. The problem isn't in the filtering approach, it's in the polarization of the source material. For example:

> I tried to get it to answer some factual questions based on government statistics on topics that left of center people go politically thermonuclear on.

You're oddly reticent to say what these supposed facts are, which is apt to make one worry about your conclusions, but I'm willing to give one:

> Wikipedia: In the US, individuals identifying themselves as Asian generally tend to score higher on IQ tests than Caucasians, who tend to score higher than Hispanics, who tend to score higher than African Americans.

Now, responses to this statement tend to "cohere" into one of two ideological bundles:

1. Denies that there are differences in racial performance in IQ tests, denies that IQ test measures anything significant, denies that differences in IQ test performance are the result of genetic differences, denies that difference in intelligence explains criminality, denies that difference in intelligence explains or justifies differences in economic success and wealth.

2. Asserts that there are differences in racial performance in IQ tests, asserts that IQ tests measure "intelligence", asserts that differences in IQ test performance are the result of genetic differences, asserts that difference in intelligence explains criminality, asserts that difference in intelligence explains or justifies differences in economic success and wealth.

This coalescing of views is the result of underlying ideological commitments; on one hand, to the idea that racial inequality in the United States is a blameworthy thing that should be addressed, and on the other, to the idea that racial inequality is the result of natural forces that do not need to be addressed.

The latter camp is going to border on, and in many cases contain, overt racism. If you're in a position where you have to filter the material being used to train your chat bot, the only plausible option is to throw out the racist parts of the source material. This is inevitably going to lead you to factual biases. There's no need to result to Mao quotes to explain what is happening here.

The unfortunate reality, of course, is that facts are more complicated than ideology. (It is a fact that there are racial differences in IQ test performance. It is NOT widely accepted that these differences are due to genetic makeup or that they explain crime rates.) But unless you want OpenAI to start picking and choosing individual facts to include in their training data, I don't see how they can do better than including everything, with a filter for racism, homophobia, etc.

Re:I gave ChatGPT the 117 question, eight dimensional Politi... by anonyfox on Tuesday December 06, 2022 @08:07AM (#33879045)
From a european perspective, GPT is right at the center of the political spectrum.

It might only appear leftist if you're from the US, where you vote either for the extreme rightwing or slightly rightwing party.

Re:I gave ChatGPT the 117 question, eight dimensional Politi... by nickpp on Tuesday December 06, 2022 @08:18AM (#33879194)
Please check your bias. I am also European and I find both US parties on the left side of the spectrum: big government, anti-individual freedoms, pro-labor, anti-business and pro planned economy.
Re:I gave ChatGPT the 117 question, eight dimensional Politi... by bmicraft on Tuesday December 06, 2022 @08:27AM (#33879301)
I find it hard to believe you actually think either party is pro-labor, when half the states have at will employment and the US has one of the lowest pto requirements.
Re:I gave ChatGPT the 117 question, eight dimensional Politi... by nickpp on Tuesday December 06, 2022 @08:32AM (#33879365)
See Biden's pro-union stance to the point he avoids naming Tesla when talking about EV manufacturers because Tesla is anti-union.
Re:I gave ChatGPT the 117 question, eight dimensional Politi... by wlesieutre on Tuesday December 06, 2022 @10:20AM (#33881406)
Biden is so pro-union that when a union threatens to strike for paid sick time he'll tell them to suck it up and sign a law forcing them to work
Re:I gave ChatGPT the 117 question, eight dimensional Politi... by nickpp on Tuesday December 06, 2022 @01:12PM (#33884552)
Maybe he didn't want the economy shut down on his watch. But if it took the most powerful man in the world to stop them, US Unions are powerful indeed.
Re:I gave ChatGPT the 117 question, eight dimensional Politi... by wlesieutre on Tuesday December 06, 2022 @02:45PM (#33885946)
If he’s so big government, anti-individual freedoms, pro-labor, anti-business and pro planned economy, then surely he could wave his president wand and force the rail companies to concede sick days instead of letting them shut down the economy
Re:I gave ChatGPT the 117 question, eight dimensional Politi... by nickpp on Tuesday December 06, 2022 @03:02PM (#33886221)
Touché. But a deal was imposed, just not the requested paid sick days. With such massive government interventions though, this sounds like anything but a free-market environment.
Re:I gave ChatGPT the 117 question, eight dimensional Politi... by bmicraft on Tuesday December 06, 2022 @08:32AM (#33879363)
From you comment history, you really love talking about a "planned economy". I wonder what that means to you. Usually it means the government does the planning, but that doesn't make sense in context.
Re:I gave ChatGPT the 117 question, eight dimensional Politi... by wlesieutre on Tuesday December 06, 2022 @08:26AM (#33879285)
> pro-labor

Funniest thing I've read all week, thanks

Re:I gave ChatGPT the 117 question, eight dimensional Politi... by endymi0n on Tuesday December 06, 2022 @08:33AM (#33879375)
> pro-labor

True that, mate. In no developed country you can do more unskilled labor for the same money.

Re:I gave ChatGPT the 117 question, eight dimensional Politi... by bartread on Tuesday December 06, 2022 @08:28AM (#33879305)
> Please control your bias.

I don't really want to get deep into this, but both you and GP must see the irony in your positions?

Re:I gave ChatGPT the 117 question, eight dimensional Politi... by Mordisquitos on Tuesday December 06, 2022 @09:02AM (#33879857)
Do you really find them “on the left side of the spectrum”, or do you actually find them “to the left of your personal beliefs”? Those are not the same thing.

It is clear that GP commenter was comparing the positions of the U.S. Democratic and Republican parties to the positions of most mainstream parties in Europe. Have you got any examples in which the overall Democratic (let alone Republican) party policy would be considered left wing as compared to that of the parties in any European parliament or assembly?

Re:I gave ChatGPT the 117 question, eight dimensional Politi... by GameOfFrowns on Tuesday December 06, 2022 @09:23AM (#33880261)
>From a european perspective, GPT is right at the center of the political spectrum.

From a European perspective many talking points and political efforts of the Democratic party are certainly extremely left-wing. In fact, many of the European left-wing parties crudely import talking points on internal US issues, like for instance Black Lives Matter. Also, why would a "slightly rightwing party" push for anti-white CRT in schools, sex transitioning of children, abortions and mass immigration?

Re:I gave ChatGPT the 117 question, eight dimensional Politi... by ShellfishMeme on Tuesday December 06, 2022 @10:22AM (#33881449)
Your last sentence makes it hard to take any ideas you have about where on the political spectrum anything lies seriously.
Re:I gave ChatGPT the 117 question, eight dimensional Politi... by dragonwriter on Tuesday December 06, 2022 @01:38PM (#33884916)
> The flaw isn't that it's biased, it's just that its creators say that it is unbiased

Where, precisely, does OpenAI claim that ChatGPT is unbiased?

ChatGPT obviously has biases from its training data, and a separate set of biases from the filters OpenAI has (not at all secretly) placed over it to avoid content that OpenAI doesn’t want it to return.