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

Who Should Stop Unethical A.I.?

newyorker.com

Post reply on HN
Print view View original

91–100 of 163 posts

Re: Who Should Stop Unethical A.I.?

#91

Post by IfOnlyYouKnew »

iujjkfjdkkdkf wrote: post #66
iujjkfjdkkdkf wrote: post #12

The more I think about this, the more inclined I am to see this kind of article as a kind of "fake news" that is deliberately (or with willful negligence) misrepresenting facts. I need to be clear that I don't believe in censorship and it's fine if this is what the author wants to write, but it should be called out how irresponsible it is. On a place like HN, the audience is generally able to read such things critica…

Let me guess, it was saying I dont like censorship that set people off :)

No, it was where you are claiming that the article contains deliberate (“fake news”) lies when it doesn’t. Because, among other reasons, it’s mostly quoting people’s opinions.

Re: Who Should Stop Unethical A.I.?

#93

Post by IfOnlyYouKnew »

A12-B wrote: post #83
AngrySkillzz wrote: post #77

People are already contributing code and resources to the actual catastrophic, existential AI risk du jour, the one that is likely to cause massive political disruption and potentially attendant destruction of value and/or human life: the algorithmic news feed.

It's probably already done that. My bet on what the next AI disaster will be is a doctored video and audio, erasing people's trust in evidence and ruining the court system, more or less. You could even argue witnesses are no longer good evidence since everyone is being manipulated by AI news feeds ;)

Very little reported news relies on video, and even less on video that comes from outside a chain of known entities (I e. CNN broadcasting something they filmed themselves).

People getting their news from random videos posted on Facebook are already lost to democracy for all practical purposes, so little would really change.

Re: Who Should Stop Unethical A.I.?

#95

Post by whakim »

Even with greater awareness of the biases inherent in e.g., training data, better and better models (and larger and better datasets) are going to lead to AI applied to ethical grey areas. This article doesn't discuss explainability and contestability, which to me seem like critical areas we need to be putting more thought into given the ultimate inevitability of biased and unethical AI.

Re: Who Should Stop Unethical A.I.?

#96

Post by lordCarbonFiber »

elihu wrote: post #68
chairmanwow1 wrote: post #4

I find Alex’s comments pretty openly inflammatory and sensationalist. It’s a pretty far throw from a system that predicts a face from a voice to transphobia.

I agree to the extent that labeling something as "transphobic" is a stretch when the researchers probably didn't even think about transgender people at all. Maybe "transinconsiderate" would be a more accurate label? Or Alex could have just said "this research reinforces cultural stereotypes and prejudices in ways that could cause problems for people like me sometime down the line". On the other hand, I think it's leg…

A mistake that a lot of people make when talking about bigotry is assuming it begins and ends with loud angry people yelling slurs. Building a system that genders people based on their voice isn't "inconsiderate" it's just a transphobic thing to do. Systematic biases don't require malice from any individual working within them to perpetuate; but that doesn't make them any less real and doesn't mean that we shouldn't just accept that all tech should also support them.

Re: Who Should Stop Unethical A.I.?

#97

Post by klyrs »

refenestrator wrote: post #67
klyrs wrote: post #63

Earlier quoted context omitted.

> At what point does this stop being AI's fault and start being an accurate observation of things that are society's fault? AI shouldn't take the blame. Blame the folks collecting biased data, or those making biased decisions encoded in the data. The data is known to be tainted. Blame those using that data to train models, and sell/rent/apply those models for profit. Blame the researchers who know, or should know, be…

But the underlying situation is biased. The data could be both accurate and unfair. Is it, just don't do data, same interest rate for everyone, no denials and amortize defaults across higher rates for lower-risk borrowers? You'd need a law, the first bank to do that would be crushed by other banks that can attract the lower-risk borrowers with lower rates.

> You'd need a law, the first bank to do that would be crushed by other banks that can attract the lower-risk borrowers with lower rates.

Curious. Is your answer to the title "we need laws to regulate AI?" I'm not specifically agreeing or disagreeing.

Re: Who Should Stop Unethical A.I.?

#98

Post by SpicyLemonZest »

joshuamorton wrote: post #88
SpicyLemonZest wrote: post #80

Earlier quoted context omitted.

I'm very skeptical that this is a sufficient replacement. Relevant expertise matters. If all the top economics conferences adopted a code of ethics saying they can't study minimum wages or tax increases, that would be a serious constraint on our ability to understand the issues, even if think tanks and sociology conferences ignored those constraints. Perhaps equally importantly, I'm skeptical that you can limit ethic…

So let's examine our axioms: I'd, in some ways, expect Dr. Hanna to ascribe to the axiom that predictive policing is unethical. As such, applied research into "improving" predictive policing would also sit squarely in the unethical space. And I expect that's all you're going to get at an ML conference (outside of explicitly ethics/critique papers): "Our model predicts crime 4.2% better than competing models" However,…

But I just don't see how you can meaningfully study the ethics of predictive policing without building potential predictive policing models and studying them for flaws. How can we talk in generalities about the ethics of predictive policing if we don't know how it works or what its effects are? I've seen articles that try to, and they just get dragged into the larger political debate about whether police are good in general.

Re: Who Should Stop Unethical A.I.?

#99

Post by A12-B »

efitz wrote: post #86
A12-B wrote: post #83

Earlier quoted context omitted.

It's probably already done that. My bet on what the next AI disaster will be is a doctored video and audio, erasing people's trust in evidence and ruining the court system, more or less. You could even argue witnesses are no longer good evidence since everyone is being manipulated by AI news feeds ;)

I’m not worried about that at wrt courts; they already know how to handle hearsay and log files and other digital evidence subject to tampering. I’m worried about social media and news wrt deep fakes. With today’s media climate I can see this being exploited to trash someone’s reputation or spin a narrative way faster than anyone would be able to reliably prove the fakery. Imagine someone faking a video of you using…

I don't think this is as much of a problem as the majority of hackernews seems to think. Of course, that is controversial to say here, but I was told I shouldn't be conventional minded.

At the end of the day the world is shaped more by material conditions than ideas. That doesn't excuse the cases you're worried about, but I think the majority of violently polarising media will only end up being so because there was already some resource dispute in the first place. It would make more sense, then, to tackle those issues, rather than fruitlessly trying to engineer the human brain to be less susceptible to rhetoric.

Re: Who Should Stop Unethical A.I.?

#100

Post by nradov »

AngrySkillzz wrote: post #77

People are already contributing code and resources to the actual catastrophic, existential AI risk du jour, the one that is likely to cause massive political disruption and potentially attendant destruction of value and/or human life: the algorithmic news feed.

Do the various social media news feeds actually use a significant amount of AI? I had understood they mostly used simpler, less computationally expensive statistical algorithms.
Post reply on HN