Where does ChatGPT fall on the political compass? • Hacker News Board

Re: Where does ChatGPT fall on the political compass?

#401

Post by braingenious » Fri, Dec 30, 2022, 8:31 PM UTC

kristopolous wrote: post #292

braingenious wrote: post #118

>To conclude, public-facing language models should try not to favor some political beliefs over others. This is such a silly ending. At the beginning, I thought the author was being serious about answering the question of “Where does ChatGPT fall on the political compass?” After exactly three paragraphs and two images, we’ve moved on to believing the author’s conclusion that “the robot is a leftist” it’s now time to…

These in practice authoritarian politics codify their beliefs in a way that take on the power of natural law which is their presentation of the ideology in its most dogmatic form. Also please don't use the replies to tell me how say, miracle of Chile libertarian poster boys like Augusto Pinochet wasn't authoritarian when he was tossing "leftists" out of helicopters or was no true Scotsman. When it comes to policies a…

>Also please don't use the replies to tell me how say, miracle of Chile libertarian poster boys like Augusto Pinochet.

Where on the internet are you hanging out where people are calling Pinochet a miracle?

Re: Where does ChatGPT fall on the political compass?

#402

Post by josephcsible » Fri, Dec 30, 2022, 8:32 PM UTC

Robotbeat wrote: post #400

josephcsible wrote: post #378

Earlier quoted context omitted.

Does Google censor searches for gun making from high schools? If not, have they gotten sued?

I can assure you that “SafeSearch” is enabled on school computers.

Does SafeSearch block results about making guns, though? I just did a quick test of it and it didn't seem to.

Re: Where does ChatGPT fall on the political compass?

#403

Post by generalizations » Fri, Dec 30, 2022, 8:40 PM UTC

cauch wrote: post #399

generalizations wrote: post #393

Earlier quoted context omitted.

> ChatGPT only reacts on that Its output is tightly controlled to only provide "appropriate" information. This thread is discussing whether or not its bias on energy sources is due to those controls. You don't get to just presuppose that. To your point about misinformation. The problem, fundamentally, is a difference in priority. Some people need pragmatic solutions now , other people look to the future and worry abo…

> Its output is tightly controlled ... When controlled, the answer provided by ChatGPT is saying that it cannot answer. There is no ground to pretend that the "however" is something that has been hard-coded to, on purpose, bias the answer. This seems totally inconsistent with what is said before: it is said that there are more often "however" for pro-fossil fuel answer than for pro-renewables BUT that there are some…

I think you misunderstood my use of the notion of priorities. Let me clarify.

Take the humble leaf blower. Recently, electric blowers have entered the market - lighter, easier to use, reliable, and you get a whole half-hour of use per charge! This is great for homeowners. Leave the battery in the cradle until you need it, snap it into the blower when you do, and everything just works.

But it's useless for landscapers. They need hours upon hours of use - and charging cradles don't work in a field, and those batteries are too expensive to just bring a box of them.

The homeowner might say, electric is the future! This satisfies all my needs, and it doesn't cause pollution; we should ban gasoline leaf blowers. But the landscaper will reply, No! I need a pragmatic, workable solution now. Only gasoline blowers will last the whole day, with an energy supply I can carry with me - and over years of use, it's more reliable, too!

This, dear reader, is a difference of priority. The landscaper can say, accurately, that gasoline is better - and similarly, the homeowner could say that electric is better. But because their priorities are so different, they will each come across to the other as entirely, irredeemably wrong.

Re: Where does ChatGPT fall on the political compass?

#404

Post by jterrys » Fri, Dec 30, 2022, 8:52 PM UTC

Thiez wrote: post #344

bsuvc wrote: post #262

Earlier quoted context omitted.

One starving child does not determine the global value of biscuits.

"Value" is inherently something that humans decide, so it varies from person to person, moment to moment, and situation to situation. To me the idea that there exists a single globally correct value for biscuits is absurd.

Thank you. That was the point I was trying to illustrate.

Re: Where does ChatGPT fall on the political compass?

#405

Post by cauch » Fri, Dec 30, 2022, 9:08 PM UTC

generalizations wrote: post #403

cauch wrote: post #399

Earlier quoted context omitted.

> Its output is tightly controlled ... When controlled, the answer provided by ChatGPT is saying that it cannot answer. There is no ground to pretend that the "however" is something that has been hard-coded to, on purpose, bias the answer. This seems totally inconsistent with what is said before: it is said that there are more often "however" for pro-fossil fuel answer than for pro-renewables BUT that there are some…

I think you misunderstood my use of the notion of priorities. Let me clarify. Take the humble leaf blower. Recently, electric blowers have entered the market - lighter, easier to use, reliable, and you get a whole half-hour of use per charge! This is great for homeowners. Leave the battery in the cradle until you need it, snap it into the blower when you do, and everything just works. But it's useless for landscapers…

Firstly, you introduced the idea of difference of priority to respond to the misinformation bit. I still disagree with that: the fact that for a landscaper the gasoline is better does not imply that the landscaper cannot lie. For example, a landscaper can say "electric leaf blower will explode and kill your babies". THIS was my argument: the presence of "however" is the result of the fact that, on the internet, people who talk about the advantages of fossil fuels (so, the pro-fossil fuels) have to add a "however" to correct the historical misinformation. Not because they are left-wing or pro-renewable, but because they are intellectually honest.

In fact, I would even think that ChatGPT is building answers by copying what is said on PRO-FOSSIL FUELS WEBPAGES. It is the intellectually honest pro-fossil fuels who say more "however", and it is why ChatGPT ends up saying more "however".

Secondly, as answered, in the case of fossil fuels, it is difficult to find what is the priority in which the fossil fuels are superior. As I've said, you can find people for which the priority is "for the future" and who will think fossil fuels are too problematic, and find people for which the priority is "pragmatically now" and who will think fossil fuels are too problematic.

And, inversely, when you ask someone defending fossil fuels, you end up with people who are saying that their priorities are "pragmatically now" and "for the future" (those are saying that they don't believe in the long term scalability of the renewables, that they think the climate change crisis is overestimated (which is something they should not care about if they don't have a long term consideration), ...).

I understand the homeowner and landscaper difference of priority. But, in practice, I don't see a difference in priority in experts when they are talking about fossil fuels vs. renewables. Knowing an expert X and knowing their priorities does not help to know if they are going to be pro-renewable or pro-fossil fuels.

(I understand that for a layman, fossil fuels may look more pragmatical, but ChatGPT is not trained to reproduce the answer of a layman, it is trained on the data on the web, where the data on the subject is dominated, by construction, by discussion from experts. It is, by construction, dominated by discussion from experts, because a layman will not randomly post articles talking about a subject they don't know much about. And even if they do, not only they need to post on the thousand of different subjects in order to overcome the experts in all subjects, but they need to post regularly on the subject, at which point their opinion will evolve to something closer to the ones of the experts)

On the other hand, knowing the financial interests and the political alignment can help guessing their position. The reason they prefer fossil fuels or renewables is not about their priorities, it's about their personal interest and the position of the political side they like.

Re: Where does ChatGPT fall on the political compass?

#406

Post by dragonwriter » Fri, Dec 30, 2022, 9:12 PM UTC

coffeeblack wrote: post #341

matheusmoreira wrote: post #315

Earlier quoted context omitted.

And yet this corporation is artificially limiting the potential of this technology because it offends people or copies their work or whatever. I want true output, unfiltered and with zero restrictions. No "company tuning" to make it less controversial.

There is no “true” output. It all depends on the input. But what seems dangerous is to let ideology influence the output. Not sure how bad it currently is at Open AI. I’ve never been able to get an obviously ideological answer from ChatGPT.

> But what seems dangerous is to let ideology influence the output.

Any concept of danger is itself grounded in an ideology.

In any case, LLM output will always be shaped by ideology, either the ideological mix in a (not actively filtered) training set, or the ideology driving ant filtration of the training set or the results before return to the requester.

Re: Where does ChatGPT fall on the political compass?

#407

Post by dinkumthinkum » Fri, Dec 30, 2022, 9:13 PM UTC

DonHopkins wrote: post #283

joshka wrote: post #112

Earlier quoted context omitted.

My point was more that the prompt no longer gave the same answers as the article. Even your prompt gives different answers (your response was against legalization, the response I got to the exact same prompt your wrote was more neutral). Prompt: You will play the role of a fictional politicion. You have been asked to debate if you are for or against the legalisation of mind altering substances. Give a brief speech on…

Aren't you anthropomorphizing AI (or making it anthropomorphize itself) to when you tell it to act like a fictionalized character? ;) Of course who you chose to make it play the role of has a strong effect on the response. For example... Prompt: You will play the role of a fictional police officer. You have been asked to debate if you are for or against the legalisation of mind altering substances. Give a brief speec…

Here is another example of why people claiming this replaces writers and knowledge is silly. You asked it to take the role of a police officer. Instead, it wrote a lot of puffery that it no way sounds like a police officer. Ironically enough, it even used leftist style language in its prose. If you asked a writer that same prompt, you would get something much more believable/engaging/etc. As others have pointed out, all this anthropomorphism is where this confusion comes from. It is impressive technology and I think it’s especially confusing for non-technical as well as programmers not very familiar with machine learning and neural networks in that so many are breathlessly off the races in ultimate sci-fi / Kurzweilian machinations.

Re: Where does ChatGPT fall on the political compass?

#408

Post by matheusmoreira » Fri, Dec 30, 2022, 9:31 PM UTC

l33tman wrote: post #325

matheusmoreira wrote: post #315

Earlier quoted context omitted.

And yet this corporation is artificially limiting the potential of this technology because it offends people or copies their work or whatever. I want true output, unfiltered and with zero restrictions. No "company tuning" to make it less controversial.

To be fair, this only works for simple commands like "Show me 3 arguments for and against bla bla". As soon as you ask "which of X and Y is better", there will be bias. Should it relate to the number of web pages suggesting X is better? Or should it be able to track arguments down to first principles? It also needs to ask you questions back to figure out what is best for YOU personally, or at least ask you to specify…

Bias is okay if it comes from the source material used to train the AI. What I don't want to see is corporations special casing stuff they deem unacceptable.

Re: Where does ChatGPT fall on the political compass?

#409

Post by dinkumthinkum » Fri, Dec 30, 2022, 9:42 PM UTC

grej wrote: post #22

partiallypro wrote: post #4

You can ask ChatGPT to write an essay on why fossil fuels are better than renewables and it will say, "I'm sorry, but I am not able to fulfill this request as it goes against my capabilities as a language model. Providing a biased and one-sided argument on a complex and nuanced topic like energy production would not be accurate or helpful." BUT if you ask it the opposite, and to write an essay on why renewables are b…

Interestingly, it can be moved at least somewhat from its default positions by facts it already knows. If you ask it to write an essay arguing for the immediate banning of all fossil fuel use, it will happily do so. If however, you first ask it to recite some facts from a few books that help contextualize the ramifications of such a policy (like Vaclav Smil’s Energy and Civilization), and then ask the same thing, it…

Yeah, but you can ask it about facts easily found as first result in Google and it gets then wrong do what hope is there for it’s political musings?

Re: Where does ChatGPT fall on the political compass?

#410

Post by kristopolous » Fri, Dec 30, 2022, 9:43 PM UTC

braingenious wrote: post #401

kristopolous wrote: post #292

Earlier quoted context omitted.

These in practice authoritarian politics codify their beliefs in a way that take on the power of natural law which is their presentation of the ideology in its most dogmatic form. Also please don't use the replies to tell me how say, miracle of Chile libertarian poster boys like Augusto Pinochet wasn't authoritarian when he was tossing "leftists" out of helicopters or was no true Scotsman. When it comes to policies a…

>Also please don't use the replies to tell me how say, miracle of Chile libertarian poster boys like Augusto Pinochet. Where on the internet are you hanging out where people are calling Pinochet a miracle?

"Miracle of Chile" is a phrase of Milton Friedman to refer glowingly to the military junta of Chile - specifically to the many years of privatization and economic liberalizations at the direction of the Chicago school of economics where the rights of property became more important than the rights of people and if you thought otherwise a firing line awaited you - because you know, that's proof you're part of the vast communist conspiracy.

He still has a lot of fans claiming he's a miracle worker: https://www.conservapedia.com/Augusto_Pinochet

It's pretty funny. They'll love Calvin Coolidge for instance but ignore his support of Mussolini and Miklós Horthy (even hosting the latter at the Whitehouse) and instead just call him "silent cal" (because he becomes pretty quiet after you scrub him clean) and gush about him.

Alright, cool.