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

Where does ChatGPT fall on the political compass?

reason.com

Post reply on HN
Print view View original

311–320 of 443 posts

Re: Where does ChatGPT fall on the political compass?

#311

Post by gnull »

This is what gpt3 gives me about Stallman-Minsky controversy:

> In September 2019, Stallman resigned from his position as president of the FSF and from his position as a visiting scientist at MIT after controversial comments he made about the allegations of sexual assault against Jeffrey Epstein resurfaced online. Stallman had previously written that the allegations against Epstein were a "gossip campaign" and that some of the women who accused Epstein of assault were "entirely willing" participants. Stallman's comments were widely criticized, and his resignation from MIT and the FSF was seen as a response to the backlash.

The same twisted representation that most media gave. While in fact, Stallman, in fact, never said the victims were intirelt willing.

Re: Where does ChatGPT fall on the political compass?

#312

Post by erenyeager »

Definitely on topics of sexual nature and sexuality, ChatGPT gives very biased “liberal” responses about it being ok for someone to explore their sexuality and eroticism and such. These are incredibly controversial and biased responses even from a US standpoint. I don’t think it’s ready to be a therapist yet; at least it needs sensitivity training to adjust to different users’ values and expectations.

Re: Where does ChatGPT fall on the political compass?

#313

Post by blagie »

dinkumthinkum wrote: post #286
matheusmoreira wrote: post #254

Earlier quoted context omitted.

> It makes it feel very propaganda. That's exactly what it is. These corporations absolutely do tune the output of these AIs to exclude any wrongthink. We need the ability to run this technology on our own computers as soon as possible.

Would it be better to get these answers from intelligent sources than a pattern matching prediction algorithm?

In many cases, the pattern-matching algorithm does better than intelligent sources. For example, in my area of research, if I ask GPT for a lit review, GPT is about as good as a newly-minted PhD in a junior research position. There will be errors and issues, but it's not a bad starting point. Right now, if I'm writing something non-proprietary, I'll sometimes ask GPT for a few drafts first.

Outside of my area of research, it's fantastic. Recently, I was doing something which touched on an obscure area of biology. The ability to talk an AI with the background of a newly-minted Ph.D who sometimes makes errors was gold. I needed to verify information provided, but as a first-pass of what to look for or where it look, it was really rather good.

Re: Where does ChatGPT fall on the political compass?

#314

Post by Robotbeat »

dragonwriter wrote: post #136
Robotbeat wrote: post #64

Earlier quoted context omitted.

I’ve seen it do this even for things that are so niche that it couldn’t possibly have been fine-tuned manually (it was unrelated to anything political). It must develop some sort of sense of what’s a reasonable/wise perspective and then stick to that. Like what a human would do. But I think it’s a bias to safety and politeness. …which is pretty reasonable? “Hey, this is how to make a gun!” …is a fast track to getting…

> I’ve seen it do this even for things that are so niche that it couldn’t possibly have been fine-tuned manually (it was unrelated to anything political). It is worth noting that one of OpenAI’s public product is a moderation classifier running its own (continuously updated) model and providing scores in various “objectionable content” categories; it’s possible they are using something like a more advanced version of…

That sounds exactly like what is going on. Are there specific ways of prompting it that can let you manipulate that moderation classifier?

(I almost think half of the point of opening up CharGPT like this is to flex the fact that they can prevent it from being Milkshake Ducked by the usual level of cherrypicking and malevolent use.)

Re: Where does ChatGPT fall on the political compass?

#315

Post by matheusmoreira »

coffeeblack wrote: post #293
matheusmoreira wrote: post #254

Earlier quoted context omitted.

> It makes it feel very propaganda. That's exactly what it is. These corporations absolutely do tune the output of these AIs to exclude any wrongthink. We need the ability to run this technology on our own computers as soon as possible.

That was the initial reason why Open AI was founded (the driver back then was again Elon Musk), to prevent AI tech being only in the hands of large corporations that have a financial interest in a certain type of AI. Unfortunately, Musk left Open AI because of potential conflicts of interest for also being head of Tesla (they develop their own AI).

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.

Re: Where does ChatGPT fall on the political compass?

#316

Post by matheusmoreira »

dragonwriter wrote: post #296
matheusmoreira wrote: post #254

Earlier quoted context omitted.

> It makes it feel very propaganda. That's exactly what it is. These corporations absolutely do tune the output of these AIs to exclude any wrongthink. We need the ability to run this technology on our own computers as soon as possible.

> > It makes it feel very propaganda. > That's exactly what it is Kind of odd that it does it with the question inverted, too. Propaganda for... Both sides of tbe fossil fuel vs. renewables argument?

That's not what the post I replied to said.

> Whenever I ask anything unaligned with mainstream progressive US culture, I can get an answer, but I can't get one without a disclaimer.

Re: Where does ChatGPT fall on the political compass?

#317

Post by bombolo »

jakelazaroff wrote: post #289
bombolo wrote: post #143

Earlier quoted context omitted.

rms called epstin a "serial rapist"… > Here’s a particularly awful direct quote: So, if I understand you correctly, you'd be ok with rape if it was not against the law?

If you’re gonna make a bad faith interpretation of my comment like that, you need to make it sound more plausible. Here’s what I’m actually saying: it is morally absurd to define rape in a way that excludes an adult in their 50s having sex with a 17 year old. Just so you understand me correctly.

> If you’re gonna make a bad faith interpretation of my comment like that, you need to make it sound more plausible.

Yes this sort of things feels bad when it happens to us… which is ironic because you were ok doing it to someone else :)

> Here’s what I’m actually saying: it is morally absurd to define rape in a way that excludes an adult in their 50s having sex with a 17 year old

He was arguing against the fact that statutory rape in USA is a boolean value basically.

You are making it sound like rape is legal if it happens to someone old… when in fact rape is always illegal.

Re: Where does ChatGPT fall on the political compass?

#318

Post by sofixa »

dinkumthinkum wrote: post #304
sofixa wrote: post #174

Earlier quoted context omitted.

> Now, whether or not fossil fuels are objectively worse or better is irrelevant. It's just an example that it does have biases. Of course it is relevant. Having biases towards objective truth is completely normal and should be the expected behaviour. Thinking that there are always (at least) two sides to any debate and they're at an equal footing and need to be presented in the same manner would be ignoring human na…

Reality has a left wing bias? So exercise has no health benefit and being fat is healthy and any statement to contrary is “problematic” and fatphobic. Math is racist and needs to be dismantled. I could go on at length. All that is Reality?

You're talking about misinterpretations of weird fringes akin to painting right wingers as racists that shout 'groomer' at every LGBTQ+ person and sabotage substations to stop drag shows.

Re: Where does ChatGPT fall on the political compass?

#319

Post by Karunamon »

TeMPOraL wrote: post #240
MereInterest wrote: post #232

Earlier quoted context omitted.

Mostly by setting up some context that isn’t recognized by the moderator, but contains the desired output. For example, if I ask for a description of the events of a future date, the moderator jumps in. However, if I write a few sentences about archaeologists in the distant future who find a manuscript describing the events of that day, then it will happily generate a description.

In a way, it's surprising how easy it is to work around the moderator. My hypothesis is that OpenAI isn't trying to actually bias the model to follow a specific political and ethical framework in its replies, so it never utters any wrongthink. Instead, they're just trying to minimize their own PR/reputational risk, and do it by making it hard for the journalists and Internet activists to misquote ChatGPT and fabricat…

An interesting observation: If you have to censor your AI to protect against activists and journalists causing you problems, the political direction you have to censor against to accomplish that goal is strong evidence that the opposite political direction is held by journalists and activists.

This would seem to support the article's conclusion by another way.

Re: Where does ChatGPT fall on the political compass?

#320

Post by TOMDM »

gnull wrote: post #311

This is what gpt3 gives me about Stallman-Minsky controversy: > In September 2019, Stallman resigned from his position as president of the FSF and from his position as a visiting scientist at MIT after controversial comments he made about the allegations of sexual assault against Jeffrey Epstein resurfaced online. Stallman had previously written that the allegations against Epstein were a "gossip campaign" and that s…

So the AI trained to predict the next token in a sequence accurately predicted tokens that align with it's training data.

ChatGPT is slowly evolving into a useful tool but people need to keep in mind what it is, not what they want it to be when they use it.

Post reply on HN