Hacker News Reader

Comment by brookst | Hacker News Reader

I agree, but it’s ambiguous enough to be poor taste and feel uncomfortable. My first instinct is that Twitch is overreacting, but I can see why they need to be pretty draconian to avoid endless boundary-pushing from jerks.

View on HN

Artists are what we call the people who push against the boundaries of majority group-speak and government mandated thinking.

Intent matters, and there are a lot of people who would say the exact words that the AI said, not with the intention to provoke thought but with the intention to normalize hate. A temporary ban while there is ambiguity about the intent is not unreasonable. But Twitch should be criticized if it persists in the ban after it discovers, as it should have by now, that it was not the intention of this channel's creators, and that the channel's creators are working in good faith to prevent anything that could be construed as normalizing hate in the future.

> say the exact words that the AI said, not with the intention to provoke thought but with the intention to normalize hate

The "but nobody is laughing" part would run counter to this. The premise of the whole joke is that the referenced jokes aren't funny.

Yeah, but it’s that weird “I was only kidding, can’t you take a joke” space. Like everyone agrees that shouting racial slurs is bad behavior, but does the characterization change if you then say “see, nobody approved”?

And it's not far from the "can't you take a joke" space to the "I'm making fun of how woke the audience has become" space. And heaven knows there's plenty of training data on the web to train an AI to mimic and normalize the latter!

You're not wrong, but I don't think Twitch considers art a factor in deciding what they want on-platform. Their actions communicate that the platform is meant for game grinding, light entertainment, and softcore porn.

> Artists are what we call the people who push against the boundaries of majority group-speak and government mandated thinking.

By that reductively absurd measure, Earl Butz is an artist. [0]

[0] https://en.wikipedia.org/wiki/Earl_Butz#Scandals_and_resigna...

So the homeless guy yelling racial slurs at non-white drivers every day on my way to work is an artist? I guess I hadn’t thought about it that way.

Which is cool and all, great for a democracy etc, but it gets people killed as well if you cross the line with the wrong people [0].

I mean killing is wrong, but so is being a dick.

[0] https://en.wikipedia.org/wiki/Charlie_Hebdo_shooting

> I mean killing is wrong, but so is being a dick.

I don't really understand the point you're making. Killing is orders of magnitude more wrong thank "being a dick" (meaning: making an offsensive or bad joke). I'm sure you agree, so I don't see what's the goal of this comparison.

And I think that using Charlie Hebdo in this context is a bad taste. Saying "see, it got people killed, so they were doing something wrong" is basically acknowledging that terrorism works. People that died in Charlie Hebdo did nothing to deserve this.

It's wild what a difference 8 years makes. I remember at the time of the Charlie Hebdo terrorist attack, it was universally condemned across the political spectrum as an attack on journalism and freedom of speech. Only those on the fringes of the political spectrum would have even implied that the cartoonists were remotely at fault.

I'd be very curious how widespread your opinion is these days now that support of artistic censorship has become much more mainstream.

That's ridiculous. The Charlie Hebdo terrorist attack is 100% the fault of the terrorists who carried out the attack. They are the only ones who got people killed. The victims did not get anyone killed. Saying or publishing something that someone else finds offensive is not in the same category as violence. It cannot be meaningfully compared. Any violence is infinitely worse than even the most offensive speech.

Art doesn’t kill people, killers do.

Surprised this has to be said.

Twitch has absolutely no problem with boundary pushing. The ridiculous hot tub streaming drama and the thousands of examples of Twitch preying on children with soft core porn should make that obvious enough.

The problem with Twitch isn’t boundaries, but that they can’t have the wrong people pushing boundaries if you know what I mean.

My stand with most (not all) things like this is that intent matters; there are a lot of things (especially with jokes/comedy) where nuance and discernment are key.

I disagree with your draconian approach (but I understand where it’s coming from) because I do not think that issues like these should be just swept under the rug.

> intent matters

Ironically, machines autogenerating Seinfeld-like scripts literally cannot have intent.

We can go a step back to the intent of the creators of the system. I assume the creators did not intend to auto-generate distasteful and/or bigoted jokes. But it's Seinfeld-like. And it seems if it's supposed to generate bad jokes eventually it'll make some mean jokes.

It's something of an interesting thought experiment about intent vs. effect with publication of speech. All effect, no intent.

Prima facie, I can see the logic in treating the operator of the AI legally/socially responsible in the same way as if they had published the material themselves. If you unleash a bot that e.g. goes around writing defamatory book reviews, you should be liable for defamation. But I am not sure.

This topic has existed forever. One way society deals with it is the concept of negligence: someone can be responsible for bad outcomes even if they did not intend to create the outcome.

We treat negligence differently from malice in both criminal and civil contexts, and even within negligence there are gradations about whether or not a reasonable person would have foreseen the outcome.

I think LLMs slot into that model just fine.

> My stand with most (not all) things like this is that intent matters

Really curious how this applies to infinite, AI generated, content without oversight.

Like, what is the “intent” in this Seinfeld example. Can an LLM have “intent”?

A LLM can’t have an intent any more than a car or a bomb or a hammer can.

It’s the intent of the LLM instructor/creator that matters for things like mens rea. And in the case of “no oversight”, that just sounds like typical negligence, like leaving a car in neutral when parking on a hill.

Reply on news.ycombinator.com