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

Is Ethical A.I. Even Possible?

nytimes.com

Post reply on HN
Print view View original

91–100 of 100 posts

Re: Is Ethical A.I. Even Possible?

#91

Post by lemagedurage »

lakis wrote: post #74
lemagedurage wrote: post #47

Earlier quoted context omitted.

Yet murder is a punishable offense and people seem just fine with that. The law is an accepted and used set of actions that are considered bad, with associated repercussions. It's a best effort and it _works_. Why would AI have to reinvent the wheel here?

No it's not. There are cases where taking a human life is considered acceptable and ethical. Killing an abductor that threatens to kill many people and negotiations have failed is considered ethical and not punishable by law. Now if you assume that murder is the case where an unlawful act happened then you've stopped querying for ethics and your question becomes what is lawful which is again extremely ill-defined.

Which is also handled by the legal system already. Instead of going back to the drawing board, people should continue with the huge work on morality defined as rules which is the legal system.

Re: Is Ethical A.I. Even Possible?

#93

Post by nkozyra »

dogma1138 wrote: post #69
nkozyra wrote: post #44

Earlier quoted context omitted.

If the hammer is trained to operate autonomously, yes.

My dishwasher is autonomous should I it be concerned about it’s ethics?

Your dishwasher makes no autonomous decisions. AI takes training data and uses algorithms to make choices. Aware of the protocols of ethics, it can reject those choices. Without ethics, it will aim for global maxima and victory in zero sum games.

So ... no.

Re: Is Ethical A.I. Even Possible?

#94

Post by sgt101 »

thrwway19033 wrote: post #48

(Throwaway in case this is crank science.) Neural nets are inspired by human neural structures. Training is in some ways similar to human learning. Genetic algorithms especially in simulated worlds are directly inspired by the biological evolution in the real world. Is there any chance that the resulting algorithms themselves (the AI) shall have any ethical rights or significance? Especially once it exceeds in comple…

Really that's more about the effect on us and the systems of society than the intrinsic nature of rights of the AI. Having things that seem human or capable of human insight and yet making them slaves may demean us all.

Re: Is Ethical A.I. Even Possible?

#95

Post by sgt101 »

mindslight wrote: post #41
jakelazaroff wrote: post #17

Earlier quoted context omitted.

I think the point is more like: consider how often a gun is abused vs. a blanket. Surely it's not a stretch to conclude that the former enables unethical acts far more than the latter.

That framing only captures half of the truth though. The quality a gun has that a blanket does not is to enable adversarial actions. So yes, the technology of the gun does enable unethical actions like point-and-click murder. But it also enables individuals and smaller groups to defend themselves against aggressors, which most people would consider ethical. I'm not disagreeing with the idea that "artifacts have polit…

Firearms didn't democratise power in the soviet States, revolutionary China or Africa.

Re: Is Ethical A.I. Even Possible?

#96

Post by richardw »

deytempo wrote: post #73
adjkant wrote: post #2

Ethical AI is hard, but so are ethical laws, ethical taxation, and ethical advertising, to name a few more ubiquitous things than AI. The real answer here is that ethics is hard, and even if you know ethics and can get some agreement, getting humans to be ethical is even harder. AI only introduces a new danger, but at the current complexity of AI I don't think its any more significant in effect than the other ethical…

I think AI is especially important for it to be used ethically because it has an unrivaled capacity to be used unethically

Past a certain point I don't think it will be used. It will use.

Re: Is Ethical A.I. Even Possible?

#97

Post by AstralStorm »

adjkant wrote: post #76
deytempo wrote: post #73

Earlier quoted context omitted.

I think AI is especially important for it to be used ethically because it has an unrivaled capacity to be used unethically

> [AI] has an unrivaled capacity to be used unethically Does it? At a personal level, I'm much more scared of a bad person with a gun and access to me than a bad person with an AI and access to me. It's more important than the ethical use of a teacup, but certainly rivaled and beaten at a micro level. At a macro level I think your argument has more ground to stand on when it comes to things like improved efficiency o…

AI is about scale. You can achieve similar terrible results by using, say, Mechanical Turk system with additional error checks. That scales too.

The other thing AI is about is cost and perhaps secrecy. Turk system requires many people who need to be fed, that puts a pretty high low bar on price. It also requires disseminating data to many agents, each of whom could leak it.

Being more accessible means more Bad Guys (unethical actors) using it.

Re: Is Ethical A.I. Even Possible?

#98

Post by AstralStorm »

lemagedurage wrote: post #89
TheOtherHobbes wrote: post #55

Earlier quoted context omitted.

Murder is the taking of a life in a manner/context which is considered bad by definition. If you consider there are many situations where the taking of a life is considered politically or economically expedient, ethically justified, or a social or cultural necessity, the ethics of life-taking become a lot less straightforward. One of the greatest potential benefits of AI is that having to define our ethics explicitly…

In Holland we have a law which states that you can drink a maximum amount of alcohol before driving. This amount is based on research of how the alcohol concentration affects a human's ability to do things like driving. This is a sensible approach. How could AI ever reach such a specific conclusion?

Using statistics and efficiency measures alone is enough to derive this conclusion. (Albeit it might derive that having almost any BAC is bad, error rate of sensors notwithstanding.) Utilitarian not deontological but result is the same.

Re: Is Ethical A.I. Even Possible?

#99

Post by adjkant »

kthejoker2 wrote: post #75
adjkant wrote: post #18

Earlier quoted context omitted.

What is unethical about a farmer using AI to figure out the best arrangement of his crops to maximize the food he produces, maintain his land quality, and minimize his environmental impact?

I mean, the short answer is if all of the constraints and inputs to a problem are known, you don't need an AI at all. There is a guaranteed optimal crop arrangement, and the model to produce it would be purely based on natural sciences only - physics, biochemistry, engineering etc. AI is only required when there is some unknown, ambiguous, adversarial, or otherwise non-existent input or constraint. AI (or indeed any…

Would weather patterns / trends not qualify a need for AI in this context? What about prediction of food demand by type for the upcoming year? Those are the big unknown inputs an AI could pretty easily morally aid to here.
Post reply on HN