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

The tragedy of the commons, AI edition

economist.com

Post reply on HN
Print view View original

101–110 of 118 posts

Re: The tragedy of the commons, AI edition

#101

Post by AnthonyMouse »

gampleman wrote: post #80
AnthonyMouse wrote: post #48

Earlier quoted context omitted.

> That deters honest complaints from poor plaintiffs. If the complaint has a high probability of success then it has a low probability of making the plaintiff pay costs. If it has a low probability of success then isn't that what we're trying to deter?

Probability of success and merit are different things. For instance if you know for a fact that you have been fired due to your race (perhaps you were outright told this is the case), but have no evidence (you weren't recording and there were no witnesses), your case has merit (it really did happen), but unless you get lucky and some evidence appears in discovery your odds of success are miniscule. I'd argue that in…

> I'd argue that in such a case preventing such a person from seeking justice would be wrong, since they were really victims and diserve redress.

If the system isn't actually going to give them redress then what benefit is there in consuming the resources of the complainant and the court only to end up with the same result? The problem you have to solve in those cases is how to make it so someone with a meritorious claim would have a high probability of winning.

Re: The tragedy of the commons, AI edition

#102

Post by theptip »

charcircuit wrote: post #96
Tarq0n wrote: post #78

Earlier quoted context omitted.

There's still inequality in someone else receiving a false positive approval from the AI while another gets rejected. Equal treatment and stochastic systems don't mix.

The law is already a stochastic system and it's designed to eventually receive a result that is correct most of the time. It doesn't try to be 100% correct. Not having it be 100% correct does not cause society to fall over.

Exactly. Not to mention that there is no reason a well-calibrated AI system cannot be Pareto-optimal over the human-only system.

For example if you are interested in bias, an AI system can be backtested; new policies applied on old data, and whatever metric for inequality you are investigating can be measured.

You can’t backtest a human.

Re: The tragedy of the commons, AI edition

#103

Post by theptip »

Gander5739 wrote: post #19
theptip wrote: post #10

Earlier quoted context omitted.

I found this article quite myopic. This is not a “tragedy of the commons”, it’s just a deficit of state capacity. The state needs to use AI to make bureaucracy instantaneous. That’s the only option, unless you ban AI for this kind of thing (not gonna happen).

So how do you propose to deal with the inevitable errors using AI would cause?

The basics likely get you there, I don’t think the modern bureaucratic state is a particularly high bar to clear:

Measure your baseline, pick good metrics, build human-vetted ground truth, then work through false positives/negatives starting with the most egregious/surprising. Start in shadow mode until the system is clearly better, however you must define that to ensure trust/buy-in.

Re: The tragedy of the commons, AI edition

#104

Post by Dylan16807 »

AnthonyMouse wrote: post #97
Dylan16807 wrote: post #71

Earlier quoted context omitted.

People are more worried about the worst case than an imaginary average case. Unless you can get insurance for the fee, significant amounts can be a big deterrance even when they're quite unlikely. 1% might not deter, but I bet it still deters more than you want. 10% will deter a lot despite those being cases that should happen. Especially because these non-experts don't know the actual odds. Their fuzzy guessing rang…

> Unless you can get insurance for the fee, significant amounts can be a big deterrance even when they're quite unlikely. But then won't experts be offering that insurance on easy to win cases left and right? > 1% might not deter, but I bet it still deters more than you want. You don't actually want to deter meritorious cases at all, but we don't have that as an option. You can either have more false negatives or mor…

> But then won't experts be offering that insurance on easy to win cases left and right?

No, I don't think so. Too much research for not enough profit.

> And that's the case even if there isn't any formal penalty for filing a losing case, because there is still a time cost.

Time is cheap when your wages are low, while penalties are not cheap.

> Having something that can tell them that is essentially required regardless

It's not a binary between knowing the odds accurately or not at all. Even with good advice there will still be uncertainty.

Re: The tragedy of the commons, AI edition

#105

Post by perching_aix »

arionhardison wrote: post #92
perching_aix wrote: post #42

Earlier quoted context omitted.

I do, though I'm not entirely sure what am I looking at on those links. Could you start by explaining that? They look like training courses or something. I saw a sequence diagram browsing around, seemed to be specific to a sample scenario?

I fixed my open protocols yesterday because they explain the flows better: https://ice.dhs.dev/protocol/13729 hope this helps.

I'll check through in a bit, had a rough Monday. Thanks for all the explanations, I appreciate it.

Re: The tragedy of the commons, AI edition

#106

Post by Gander5739 »

theptip wrote: post #103
Gander5739 wrote: post #19

Earlier quoted context omitted.

So how do you propose to deal with the inevitable errors using AI would cause?

The basics likely get you there, I don’t think the modern bureaucratic state is a particularly high bar to clear: Measure your baseline, pick good metrics, build human-vetted ground truth, then work through false positives/negatives starting with the most egregious/surprising. Start in shadow mode until the system is clearly better, however you must define that to ensure trust/buy-in.

Any appeals process naturally falls victim to the original problem, and so we're back where we started. And if you don't have an appeals process, how can you flag if something is wrong?

Re: The tragedy of the commons, AI edition

#107

Post by weakened_malloc »

j-bos wrote: post #88
weakened_malloc wrote: post #73

Earlier quoted context omitted.

Is it any different to how it was previously? Before the age of AI, companies write all sorts of bullshit stuff into ToS or EULA which was neither enforceable or legal yet they routinely got away with it because Joe Schmoe had no ability to assess their arguments or fight back. The only difference now is that the general population has gained access to low cost legal help and are able to push and argue (whether right…

Problem: the plebs are being opressed Solution: arm the oppressed New problem: civil war

So you think that legal skills / help should remain expensive?

Re: The tragedy of the commons, AI edition

#108

Post by jackvalentine »

weakened_malloc wrote: post #73
jackvalentine wrote: post #24

Even the “pre-case” legal stuff is out of control. In my Australian organisation we’ve noticed a deluge of what I can only call “privacy demands” that talk the talk… but have the jurisdiction or subject legislation wrong. But AI told them they could demand these things (even if it would be illegal for us to comply! Think KYC stuff.) and they don’t have the skill to criticise the AI or know when to stop.

Is it any different to how it was previously? Before the age of AI, companies write all sorts of bullshit stuff into ToS or EULA which was neither enforceable or legal yet they routinely got away with it because Joe Schmoe had no ability to assess their arguments or fight back. The only difference now is that the general population has gained access to low cost legal help and are able to push and argue (whether right…

> Is it any different to how it was previously? Before the age of AI, companies write all sorts of bullshit stuff into ToS or EULA which was neither enforceable or legal yet they routinely got away with it because Joe Schmoe had no ability to assess their arguments or fight back.

Well yes, because we don't do that. You might be shocked to hear we don't even have a ToS or EULA and the relevant legislation acts as that for our purposes.

Re: The tragedy of the commons, AI edition

#109

Post by j-bos »

weakened_malloc wrote: post #107
j-bos wrote: post #88

Earlier quoted context omitted.

Problem: the plebs are being opressed Solution: arm the oppressed New problem: civil war

So you think that legal skills / help should remain expensive?

"So you think {}" in the wild! =D

Ty, and to answer your question: No. I'm commenting on a neat parallel. Personally I like the improvements that come from fairer games.

Re: The tragedy of the commons, AI edition

#110

Post by charcircuit »

sjsdaiuasgdia wrote: post #87
charcircuit wrote: post #77

Earlier quoted context omitted.

AI would review the appeal so that only things which require human decision receive human attention.

And we trust the AI to reliably decide what requires human attention because...?

Because modern LLMs are capable. Even if they make a mistake by appealing you can clear it up and convince it why there actually is an argument than it being able to be resolved by just reading the contract for example.
Post reply on HN