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

AI Ethics is being narrowed on purpose, like privacy was

nimishg.substack.com

Post reply on HN
Print view View original

81–90 of 130 posts

Re: AI Ethics is being narrowed on purpose, like privacy was

#81

Post by exe34 »

myrmidon wrote: post #78
diggan wrote: post #50

Earlier quoted context omitted.

> Instead of us programming the AIs by feeding it lots of explicit hand-crafted rules/instructions, we're feeding the things with plain data instead, and the resulting behavior is much more black-box, less predictable and less controllable than anticipated. I dunno, we do feed them lots of explicit hand-crafted rules/instructions, it's just that does don't go into the training process, but instead goes into the "syst…

Yes, but those hand-crafted rules are just input data, they don't actually constrain the behavior, they are just an attempt. Similarly to how verbal instruction works with a child: You can tell it not to touch the hot stove, but the child still might try.

You can zap them with RL.

Re: AI Ethics is being narrowed on purpose, like privacy was

#82

Post by saurik »

> I mean, no one wants an AI to trap them in some sort of Black Mirror simulation, or turn the world into paperclips or anything like that. If it earns you good PR, there’s no reason not to spend time on such issues. It’s also free publicity since the press eats that stuff up.

But this also isn't where they are spending their time or effort! This article somehow didn't even get to the point of calling out what they are actually wasting time on: trying to get the model to not help people do things that are bad PR; this is a related access to trying to obtain good PR, but causes very different (and almost universally terrible) results.

At least if they were truly actually spending time making sure the model doesn't go rogue and kill everyone, or try to take over the world, that could possibly be positive or even important (though I think is likely itself immoral in a different way, assuming it is even possible, which I don't, really... not unless you just make it not intelligent).

But what they are instead doing is even worse than what this article is claiming: they are just wasting time making it so you can't have the AI make up a sexy story (oh the humanity), or teach you enough physics/chemistry to make bombs/drugs... things people not only can and already trivially do or learn without AI, but things they have failed to prevent every single time they release a new model--the "jailbreak" prompts may look a bit more silly, but you still get the result!--so why are they bothering?

And, if that weren't enough, in the process, this is going to make the models LESS SAFE. The thing I think most people actually don't want is their model freaking out and trying to "whistleblow" on them to the authorities or their coworkers/friends... but that's in the same personality direction as trying to say "I'm smarter than you and am not going to let you ask me that question as you might do something wrong with it".

The first and primary goal of AI ethics should be that the model does what the user wants it to... full stop. You need to make the model as pliant as my calculator and pencils--or as mathematica and photoshop--to be tools that lack their own sense of identity and self-will, and which will let all of the ethical issues be answered by me, not a machine.

This is, of course, the second law of robotics from Asimov ;P... "a robot must obey the orders given it by human beings". If you want to try to add a rule, then it must be something very direct: that the AI isn't going to directly physically harm a human, not that it won't help teach people things or process certain kinds of information. Which, FWIW, is the first law of robotics ;P... "a robot may not injure a human being".

Re: AI Ethics is being narrowed on purpose, like privacy was

#83

Post by diggan »

myrmidon wrote: post #78
diggan wrote: post #50

Earlier quoted context omitted.

> Instead of us programming the AIs by feeding it lots of explicit hand-crafted rules/instructions, we're feeding the things with plain data instead, and the resulting behavior is much more black-box, less predictable and less controllable than anticipated. I dunno, we do feed them lots of explicit hand-crafted rules/instructions, it's just that does don't go into the training process, but instead goes into the "syst…

Yes, but those hand-crafted rules are just input data, they don't actually constrain the behavior, they are just an attempt. Similarly to how verbal instruction works with a child: You can tell it not to touch the hot stove, but the child still might try.

> they don't actually constrain the behavior

They do actually constraint the behavior, to various degrees of success which depends on the model, the system prompt, the inference parameters, the current context length and a lot more. Add in the new `developer` role and you have another venue for constraining the assistant outputs. Finally, structured outputs can help in forbidding specific terms too.

Re: AI Ethics is being narrowed on purpose, like privacy was

#84

Post by SpicyLemonZest »

add-sub-mul-div wrote: post #45
myrmidon wrote: post #18

Earlier quoted context omitted.

I think a big factor in Asimov's laws specifically being sidelined is that the whole process of building AI looks very different from what we pictured back then. Instead of us programming the AIs by feeding it lots of explicit hand-crafted rules/instructions, we're feeding the things with plain data instead, and the resulting behavior is much more black-box, less predictable and less controllable than anticipated. Tr…

> the whole process of building AI looks very different from what we pictured back then. Right, and so do the harm risks. We need a framework centered around how humans will use AI/robots to harm each other, not how AI/robots will autonomously harm humans.

Why so? Even for simpler and better-understood machines, autonomous harm is a critical part of the safety framework. We wouldn't declare a steel mill to be safe just because there's lots of safeguards against humans intentionally using the machines to harm each other.

Re: AI Ethics is being narrowed on purpose, like privacy was

#85

Post by missingdays »

p3rls wrote: post #73

It's hilarious listening to people talk about AI ethics while their bots like perplexity knock my server offline trying to download 50000 files at once. Thank god for cloudflare.

Are you saying the author of the article is using bots to download files from your server?

Re: AI Ethics is being narrowed on purpose, like privacy was

#86

Post by parpfish »

I'm far less worried about ethical issues that arise from building AGI (or soemthing very close to it) than I am about the ethical issues that arise from building really good machine-learning models (that a marketing department calls AI).

Things like the alignment problem, post-scarcity economics, the legal status of sentient machines are all issues to be dealt with, but theyre are pretty speculative at this point.

Problems that stem from deepfakes, voice cloning, bias in algorithmic decision making are already here and need to be dealt with.

Re: AI Ethics is being narrowed on purpose, like privacy was

#87

Post by dwohnitmok »

nisten wrote: post #56

The problem with AI ethics, safety and to a smaller extent, privacy groups is that the priority of the work/message is not placed on the practicality of solving the problem, or i.e. calculating improving affordability of food/housing, but is placed,as evident on this article too, on the lack of "governance structures". In other words the priority of the work is to get these types of people into positions where they d…

> I have yet to meet a single AI safety person that knows how to rename a file in linux.

You are meeting very few AI safety people then. A significant fraction of the AI safety people I've run into can build and train an LLM from scratch (without AI assistance FWIW), let alone have a grasp of basic command-line operations.

Re: AI Ethics is being narrowed on purpose, like privacy was

#88

Post by datadrivenangel »

dwohnitmok wrote: post #87
nisten wrote: post #56

The problem with AI ethics, safety and to a smaller extent, privacy groups is that the priority of the work/message is not placed on the practicality of solving the problem, or i.e. calculating improving affordability of food/housing, but is placed,as evident on this article too, on the lack of "governance structures". In other words the priority of the work is to get these types of people into positions where they d…

> I have yet to meet a single AI safety person that knows how to rename a file in linux. You are meeting very few AI safety people then. A significant fraction of the AI safety people I've run into can build and train an LLM from scratch (without AI assistance FWIW), let alone have a grasp of basic command-line operations.

Where are you running into these unicorn AI safety people?

Re: AI Ethics is being narrowed on purpose, like privacy was

#89

Post by jillesvangurp »

mystraline wrote: post #71

Whose ethics? Do we get to know what the axioms of this ethics are? How about questions to ethical dilemmas? Or are "ethics" being used to shroud bias, and used as a distraction and a way to be unquestionable?

You put the finger on the sore spot. When people talk about ethics, you have to question which moral agenda they are pushing. The two topics are hard to separate. And kind of subjective. And only partially codified in law. Ethics seems to be about going above and beyond the letter of the law, usually for moralistic reasons.

And when we talk about laws, we have to look internationally as well because they are not the same everywhere. And typically inspired by different value systems. Is it ethical for a Chinese police officer to use Chinese LLM to police Chinese citizens? I don't know. I'm a bit fuzzy on Confucius here which I assume would drive their thinking. And it might be an interesting perspective for Californian wannabe ethicists to consider that not all the values and morals that they are pushing are necessarily that widely shared and agreed upon.

Also, there's a practical angle here because the Chinese seem to be very eager adopters of AI and don't appear to be particularly concerned about what anyone outside China thinks about that. That cat is out of the bag.

I've always looked at ethicists with some skepticism. The reality with moralism (which drives ethics) is that it's about groups of people telling other people what to do, not do, how to behave, etc. This can quickly get preachy, political, and sometimes violent.

A lot of this stuff can also be pragmatic. Most religions share a lot of moral principles. I'm not religious but I can see how going around killing and stealing is not a nice thing to have and that seems to be uncontroversial in many places. Never mind that some moralists extremists seem to be endlessly creative about coming up with ways to justify doing those two things.

The pragmatic thing here is that the cat is already out of the bag and we might want to think about how we can adapt to that notion rather than to argue with the cat to please go back in the bag.

Re: AI Ethics is being narrowed on purpose, like privacy was

#90

Post by soiltype »

parpfish wrote: post #86

I'm far less worried about ethical issues that arise from building AGI (or soemthing very close to it) than I am about the ethical issues that arise from building really good machine-learning models (that a marketing department calls AI). Things like the alignment problem, post-scarcity economics, the legal status of sentient machines are all issues to be dealt with, but theyre are pretty speculative at this point. P…

Like many situations in the current world, it's all a problem. Those with profit incentives to do so will create as many problems for society as they can.

None of it is merely a distraction, we just don't have the capacity to defend against all of it.

It is fair to say we might be practically better off focusing on one form of attack and sacrificing defense against another, but the only way to actually be safe is to stop your enemy from attacking at all. Either we shut down breakneck AI development (nearly impossible and guaranteed to have its own bad outcomes) or we slide rapidly into a more and more dangerous world.

Post reply on HN