Comment by coldbrewed | Hacker News Reader

Comment by coldbrewed | Hacker News Reader

Parent Full thread

AI is not an inert tool that you can wield; it's a crystallization of human beliefs, knowledge, and processes. It is vulnerable to biases and this vulnerability can't be neatly chopped out because it's impossible to fully remove biases from training data.

It's significant to note that these biases don't have to be overt political biases; it can be a predisposition to use virgin materials over recycled, a tendency to generate JavaScript over Julia, etc cetera. Our available datesets are images of us and thus aren't predisposed to surpass our limitations.

Given that AI models are bound to the same biases that society struggles with, which ones do we encode in the models and which ones do we mitigate? Do we try to acknowledge and mitigate those limitations or say "This is fine" and ship our current organizational and political limitations?

If we want to stop fighting the same damn wars over and over again then our machine models have to be designed to account for our biases and limitations. Do you have suggestions for alternate ways to improve our models aside from explicit error correction?

These "biases" in data are a reflection of reality. If by default an LLM refers to engineers as "he" more than "she", it's because engineers are more likely to be male than female. Most people would prefer an AI whose model of reality matches actual reality rather than one that matches what some Silicon Valley leftists think reality should be. Those biases are only a problem if you think the main purpose of an AI is to moralise.

LLMs have no way to access reality.

Reality != whatever gets mentioned the most on the internet

The "AI ethicists" aren't trying to make AI better match reality, they're trying to make it match how they think reality should be.

This is patently false. GPT-4 will have more information pertaining to North America than to Africa because there is more training data available for North America than Africa. That's not a reflection of reality; that's a reflection of the distribution of computers and wealth.

> If by default an LLM refers to engineers as "he" more than "she", it's because engineers are more likely to be male than female.

That's a great point to raise, thank you for mentioning it. This really is the crux - machine models crystallize _current_ bias and amplify it. Most programmers in the early days of computers were women, but were displaced over time. Should the training data reflect the bias of the dominance of women in the early days, the dominance of men now, or some other combination?

> Most people would prefer an AI whose model of reality matches actual reality

Training data is not reality; training data is training data. It will fundamentally lag behind reality, and it will have a bias towards past conditions and precedents.

> These "biases" in data are a reflection of reality.

You do realise, if you command an LLM to talk like a caveman it doesn't invent a time machine, travel back in time and research how cavemen actually talked; it just produces the "me hit with rock" that redditors imagine a caveman would have talked like.

Reply on news.ycombinator.com