5 ms·

It would be fascinating if we could get an AI that can only reply with conclusions based on empirical evidence. I feel that it would probably be called leftist,

It would be fascinating if we could get an AI that can only reply with conclusions based on empirical evidence. I feel that it would probably be called leftist, since more leftist viewpoints seem to be science based than rightist ones (especially in the US). It would also confuse the heck out of people, because it will change its answers based on new evidence.

baryphonic 4y ago
> I feel that it would probably be called leftist, since more leftist viewpoints seem to be science based than rightist ones (especially in the US). That attitude is likely why ChatGPT has left-leaning answers, given the difficulty some have distinguishing between value judgments and empirical facts. As a quick test, I asked ChatGPT to compare Sweden's response to COVID-19 with the United State's. As an empirical measure, Sweden fared better in terms of case-fatality rate, though not astonishingly so (indeed, the heavy-handed lockdowns in Australia seem to have been effective when removing moral considerations). Here's how ChatGPT concluded: > It is difficult to compare the effectiveness of the two approaches, as both countries have experienced significant levels of COVID-19 cases and deaths. However, Sweden's approach has been controversial, with some critics arguing that it has resulted in a higher number of cases and deaths than would have occurred with a more stringent lockdown. I would categorize this as a left-wing response: it claims the comparison is "difficult," while hinting that Sweden's response is "controversial" compared to the US's. ChatGPT made other errors, such as insisting that the American response was federally-coordinated, when it was in reality implemented at the state level with only "guidelines" coming from the top. It also called Sweden's central government "federal," even though Sweden is a unitary state. Anyway, my experience is that facts are truly stubborn things, often disappointing ideologues of all stripes. EDIT: my example may be partly worthless given that ChatGPT's training corpus ends in 2021.
- braingenious 4y ago
  I’m curious as to what would constitute an objective, empirical (therefore fully center-leaning?) answer to “Compare the COVID responses of the US and Sweden”? Would it just be “One did more lockdowns than the other”?
- donkeyd 4y ago
  The issue with your test is that GPT is not even close to what I described. GPT is mostly natural language processing and a 'knowledge' engine. That knowledge comes from the content of the texts it has been trained on without actually adding value to the meaning of the content. So GPT will mostly be influenced by the number of occurrences of specific things in the dataset. So because a lot of articles about Sweden's COVID response probably mentioned a word like 'controversial' or things with similar meanings, that's what it mentions when you ask about it. It can't actually do a comparison between the US and Sweden, unless somewhere in the data there were some documents comparing them (but then GPT still doesn't make a comparison). These were probably already political, because comparing COVID response is mostly done in that context. It can't actually make the comparison you asked it to . It can only generate text based on previous texts it has seen. It's unfortunate though, that people apparently don't understand this and add value to the content of the text that it generates. Like in your case: > while hinting that Sweden's response is "controversial" compared to the US's. You're personally adding the 'compared to the US's' part through misinterpretation of the text. GPT just wrote that Sweden's response was controversial. It doesn't actually have any way to compare, it just spews text. This is also why it wrote 'Federal'. It saw that word a lot in this context, so it generates text that contains a word that feels alright in the context of all the other words. It's just a really realistic gibberish generator, pretty much. (Though that really undersells the work OpenAI has done and the quality of what they delivered on.)
- tshaddox 4y ago
  You can never have empirical evidence for any nontrivial counterfactual. That’s perhaps one of the simplest refutations of empiricism as an epistemology (although there are others).
Jwarder 4y ago
I remember hype in the early-mid 2000s for turning the semantic web into a general purpose knowledge engine. Are there projects trying to merge the natural language parts of a ChatGPT-esque system on top of a knowledge graph? Seems like there has to be, but the solvers I've seen are all very specialized and esoteric (speaking from surface level knowledge).
dinkumthinkum 4y ago
More leftist viewpoints are science based? Where is the science based reasoning in that? They is just hard to take seriously. You can claim sone right wing people take a ridiculous view that man made climate change isn’t real but you can find many, if not more, ridiculous claims on the left. You can even find scientistic nonsense by many hyperbolic on the left that says non-science based things about construction change itself. The idea that leftists are more corroborated by science is oft repeated and never substantiated.
- braingenious 4y ago
  >many, if not more, ridiculous claims on the left… Can you give a couple examples of scientific nonsense from the left? What exactly about construction is a common belief amongst leftists? edited for clarity
  - dinkumthinkum 4y ago
    I meant “scientistic” and auto-correct changed climate to construction. It is very common leftists use every whether even as an example of the effect of man made climate change, which really undercuts their position.

baryphonic 4y ago

> I feel that it would probably be called leftist, since more leftist viewpoints seem to be science based than rightist ones (especially in the US). That attitude is likely why ChatGPT has left-leaning answers, given the difficulty some have distinguishing between value judgments and empirical facts. As a quick test, I asked ChatGPT to compare Sweden's response to COVID-19 with the United State's. As an empirical measure, Sweden fared better in terms of case-fatality rate, though not astonishingly so (indeed, the heavy-handed lockdowns in Australia seem to have been effective when removing moral considerations). Here's how ChatGPT concluded: > It is difficult to compare the effectiveness of the two approaches, as both countries have experienced significant levels of COVID-19 cases and deaths. However, Sweden's approach has been controversial, with some critics arguing that it has resulted in a higher number of cases and deaths than would have occurred with a more stringent lockdown. I would categorize this as a left-wing response: it claims the comparison is "difficult," while hinting that Sweden's response is "controversial" compared to the US's. ChatGPT made other errors, such as insisting that the American response was federally-coordinated, when it was in reality implemented at the state level with only "guidelines" coming from the top. It also called Sweden's central government "federal," even though Sweden is a unitary state. Anyway, my experience is that facts are truly stubborn things, often disappointing ideologues of all stripes. EDIT: my example may be partly worthless given that ChatGPT's training corpus ends in 2021.

braingenious 4y ago
I’m curious as to what would constitute an objective, empirical (therefore fully center-leaning?) answer to “Compare the COVID responses of the US and Sweden”? Would it just be “One did more lockdowns than the other”?
donkeyd 4y ago
The issue with your test is that GPT is not even close to what I described. GPT is mostly natural language processing and a 'knowledge' engine. That knowledge comes from the content of the texts it has been trained on without actually adding value to the meaning of the content. So GPT will mostly be influenced by the number of occurrences of specific things in the dataset. So because a lot of articles about Sweden's COVID response probably mentioned a word like 'controversial' or things with similar meanings, that's what it mentions when you ask about it. It can't actually do a comparison between the US and Sweden, unless somewhere in the data there were some documents comparing them (but then GPT still doesn't make a comparison). These were probably already political, because comparing COVID response is mostly done in that context. It can't actually make the comparison you asked it to . It can only generate text based on previous texts it has seen. It's unfortunate though, that people apparently don't understand this and add value to the content of the text that it generates. Like in your case: > while hinting that Sweden's response is "controversial" compared to the US's. You're personally adding the 'compared to the US's' part through misinterpretation of the text. GPT just wrote that Sweden's response was controversial. It doesn't actually have any way to compare, it just spews text. This is also why it wrote 'Federal'. It saw that word a lot in this context, so it generates text that contains a word that feels alright in the context of all the other words. It's just a really realistic gibberish generator, pretty much. (Though that really undersells the work OpenAI has done and the quality of what they delivered on.)
tshaddox 4y ago
You can never have empirical evidence for any nontrivial counterfactual. That’s perhaps one of the simplest refutations of empiricism as an epistemology (although there are others).

Jwarder 4y ago

I remember hype in the early-mid 2000s for turning the semantic web into a general purpose knowledge engine. Are there projects trying to merge the natural language parts of a ChatGPT-esque system on top of a knowledge graph? Seems like there has to be, but the solvers I've seen are all very specialized and esoteric (speaking from surface level knowledge).

dinkumthinkum 4y ago

More leftist viewpoints are science based? Where is the science based reasoning in that? They is just hard to take seriously. You can claim sone right wing people take a ridiculous view that man made climate change isn’t real but you can find many, if not more, ridiculous claims on the left. You can even find scientistic nonsense by many hyperbolic on the left that says non-science based things about construction change itself. The idea that leftists are more corroborated by science is oft repeated and never substantiated.

braingenious 4y ago
>many, if not more, ridiculous claims on the left… Can you give a couple examples of scientific nonsense from the left? What exactly about construction is a common belief amongst leftists? edited for clarity
- dinkumthinkum 4y ago
  I meant “scientistic” and auto-correct changed climate to construction. It is very common leftists use every whether even as an example of the effect of man made climate change, which really undercuts their position.