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

AI at Google: our principles

blog.google

Post reply on HN
Print view View original

341–350 of 418 posts

Re: AI at Google: our principles

#341

Post by ionised »

throwaway2048 wrote: post #61
EpicEng wrote: post #55

So, I'm all for giving someone the benefit of the doubt if they have a change of heart upon reconsidering an issue, but this coming after the fact rings a bit hollow to me. I think the only principle at play here is that it became a PR issue. That's fine, but let's be honest about it. Early emails between Google execs framed this project only in terms of revenue and potential PR backlash. As far as we're aware, there…

There was a similar "comming to jesus moment" with Google in China. They saw how they had to do the right thing after years of censoring and manipulating data for the chinese government, but only after they got massively hacked, blocked, and essentially forced out of china... However a good thing done for the wrong reasons is still a good thing.

> However a good thing done for the wrong reasons is still a good thing.

Yes, but we should absolutely remember what the original intention was.

Re: AI at Google: our principles

#342

Post by ionised »

throwaway2048 wrote: post #73
ManFromUranus wrote: post #69

Earlier quoted context omitted.

Hmm not really. Google is bad and they should feel bad, you're just handwaving away how bad they are because you like them. Imagine I wanted to have somebody killed and I hired a hitman to kill them and when I go to pay the hitman I accidentally wire the money to the wrong place and inadvertantly pay off the victims mortgage instead of paying the hitman. Now the victim doesn't die and gets their mortgage paid off. I'…

Beleive me when I say that I do not like Google. But what is a person or a group of people supposed to do when they have done something wrong? All they can do is stop doing it, and try to prevent doing it in the future. You can speculate about their motives, I personally beleive it to be PR motivated aswell, but what matters in the end is that it stopped happening.

That's all fine, but that person or group shouldn't expect everybody to love them afterwards. If anything, they should expect distrust and dislike.

Re: AI at Google: our principles

#343

Post by ionised »

evilmushroom wrote: post #105
EpicEng wrote: post #55

So, I'm all for giving someone the benefit of the doubt if they have a change of heart upon reconsidering an issue, but this coming after the fact rings a bit hollow to me. I think the only principle at play here is that it became a PR issue. That's fine, but let's be honest about it. Early emails between Google execs framed this project only in terms of revenue and potential PR backlash. As far as we're aware, there…

And now someone else will easily pick up the contract, and Google loses any ability at all to influence it. Perhaps they could have reduced civilian casualities more so than whoever picks it up.

That's the biggest integrity-lacking cop-out imaginable.

"If I don't do it somebody else will!"

Re: AI at Google: our principles

#344

Post by ionised »

philwelch wrote: post #224
RyanZAG wrote: post #132

Earlier quoted context omitted.

It's a good point, and if anybody thinks an arms race isn't already happening then they are very naive. While we're debating the morals of this, the Chinese, Russians and many others are trying to make it happen. For all we know, they may have already succeeded, and we're discussing the morals of it while it's already in full production a few thousand miles away. The relevant research papers, knowledge and skills are…

And I can almost guarantee that, unlike the US, the Russian and Chinese militarized AI programs are not optimized for minimizing civilian casualties.

Doesn't seem like the US one does either.

Re: AI at Google: our principles

#345

Post by repolfx »

ergothus wrote: post #39
finnthehuman wrote: post #18

>2. Avoid creating or reinforcing unfair bias. They DO realize that the YouTube recommendation algorithm is a political bias reinforcement machine, right? Like, I think it’s fun to talk trash on google because they’re in an increably powerful position, but this one isn’t even banter.

FWIW, since they updated the google news algorithm, I've started getting a pretty diverse set of results. It might be because it picked up that I try to read opposing views, or it might be because the algorithm is terrible, or any other number of reasons, but regardless of the source it's refreshing to see. ...and revealing! The headlines differences between, say, Washington Post, CNN, The Hill, and Fox News for the…

When did that happen? I'd given up on Google News, it reached a point where it only wanted to show me articles from the Guardian or CNN even though I read a wide range of outlets. I just went back and looked again and there does indeed seem to be a better mix of stories now, although I am still very skeptical - the News algorithm is even less documented than the web search ranking algorithms and it would be tough to rebuild the trust. Google is clearly run by people who feel they need to "make the world a better place" through subtle conditioning of their algorithm outputs these days... very different to how it used to be. I'm not sure I trust them to rank news, even if improved.

Re: AI at Google: our principles

#346

Post by nightfreeze »

t1lthesky wrote: post #333
mattlondon wrote: post #328

Earlier quoted context omitted.

> It's far better for the world that the United States developed atomic weapons before either Germany or the Soviets did. [citation needed] The US having nuclear weapons didn't work out so well for the 70,000-120,000 innocent civilians that were killed in the attack on Hiroshima.[1] I don't have handy access for how many innocent civilians were killed in Nagasaki but I would assume it was similar. Would the Nazis hav…

Come on are you really making this argument? That we cant know if it was better for the USA or Nazi Germany to have nuclear weapons? Like, say you have to go back in time and give nuclear weapons to Nazi Germany or the USA during ww2.. you would throw up your hands and say we can't know who the weapons should go to?

If I understand correctly, you're saying that Germany and Russia shouldn't have had nuclear weapons because they would have used them, while arguing that it was right that they went to the Americans, in spite of the fact that they used them.

This does not seem like a valid position.

Re: AI at Google: our principles

#347

Post by repolfx »

cromwellian wrote: post #267

Several comments don't seem to understand what the "unfair bias" mentioned is. It doesn't have anything to do with censoring your favorite conservative search result. The machine learning "bias", at least the low hanging fruit, is learning things like "doctor == male", or "black face = gorilla". How fair is it that facial recognition or photo algorithms are trained on datasets of white faces or not tested for adversa…

It doesn't have anything to do with censoring your favorite conservative search result.

But that's not correct. That's exactly what it's to do with.

A big part of the philosophy of conservatism is to accept the world as you find it. Conservatives, at least in theory (not saying the Republicans are a great implementation of the philosophy), eschew large social engineering schemes, they eschew attempts to remould attitudes via manipulation of language and so on. These are all traits associated with the opposing end of the political spectrum. Think how important re-engineering people's thinking via language was in Orwell's 1984, for example.

So now we have Google and related AI researchers announcing that when an ML model learns the actual fact that most doctors are male, this is "bias" and it needs to be "corrected". This is Orwellian. It's not at all biased, it's accurate. But because of some vague, handwavy assumption that if AIs believe gender ratios are 1:1 in medicine then somehow ... via some obscure chain of action ... more girls will become doctors, the basic reality of the matter is justified in being inverted. Or possibly, more boys will choose NOT to be doctors. Quite why this outcome is better than the existing status quo is never explained with any rigour - it's ideology.

This is the very style of social engineering that conservatism rejects, on the grounds that it so often goes wrong or has unintended side effects. So whilst I am interested to see that Google is deciding to walk the walk here when it comes to AI and weapons, I nonetheless find their statement of moral principles to be quite chilling and frankly, it renders their most important products useless for me. Their take on this isn't news, but it's sad to see them doubling down on it. I do not wish to be subtly manipulated via AI driven translation or re-rankings into not believing things about the world that are true but upset radical feminists.

It's especially sad for me because I worked there for so many years.

Re: AI at Google: our principles

#348

Post by bitCromwell »

themacguffinman wrote: post #65
EpicEng wrote: post #55

So, I'm all for giving someone the benefit of the doubt if they have a change of heart upon reconsidering an issue, but this coming after the fact rings a bit hollow to me. I think the only principle at play here is that it became a PR issue. That's fine, but let's be honest about it. Early emails between Google execs framed this project only in terms of revenue and potential PR backlash. As far as we're aware, there…

I doubt that Google spelling out their moral stance is intended to convince you right away that they're all good now. It's a public standard that they're setting for themselves. If you think their actions don't match their words, you now have concrete terms and principles to critique and compare with. It's a benchmark to which employees and the public can hold them accountable.

Yes, like “do no evil”.

Re: AI at Google: our principles

#349

Post by AsyncAwait »

themacguffinman wrote: post #65
EpicEng wrote: post #55

So, I'm all for giving someone the benefit of the doubt if they have a change of heart upon reconsidering an issue, but this coming after the fact rings a bit hollow to me. I think the only principle at play here is that it became a PR issue. That's fine, but let's be honest about it. Early emails between Google execs framed this project only in terms of revenue and potential PR backlash. As far as we're aware, there…

I doubt that Google spelling out their moral stance is intended to convince you right away that they're all good now. It's a public standard that they're setting for themselves. If you think their actions don't match their words, you now have concrete terms and principles to critique and compare with. It's a benchmark to which employees and the public can hold them accountable.

Even if they abide by this, who's to say that once a party has some Google developed military AI, they won't misuse it? I fail to see how Google can effectively prevent this.

Re: AI at Google: our principles

#350

Post by mattlondon »

t1lthesky wrote: post #333
mattlondon wrote: post #328

Earlier quoted context omitted.

> It's far better for the world that the United States developed atomic weapons before either Germany or the Soviets did. [citation needed] The US having nuclear weapons didn't work out so well for the 70,000-120,000 innocent civilians that were killed in the attack on Hiroshima.[1] I don't have handy access for how many innocent civilians were killed in Nagasaki but I would assume it was similar. Would the Nazis hav…

Come on are you really making this argument? That we cant know if it was better for the USA or Nazi Germany to have nuclear weapons? Like, say you have to go back in time and give nuclear weapons to Nazi Germany or the USA during ww2.. you would throw up your hands and say we can't know who the weapons should go to?

I am saying that we cant know what Nazi Germany would have done with them - i.e. we cant know if they'd have been "good" or "bad" with them. Anything we think now is merely conjecture and speculation. We have to take a step back and really examine and question our own beliefs and biases. How much of what I am thinking is actually legit, and how much is affected by what I've seen at the movies/been told by teachers/read in the papers/seen on TV and just accepted as fact?

But we do know for an absolute fact is that the US did use nuclear weapnos to kill thousands and thousands and thousands of innocent men, women and children (and for the sake of balance, we do know for an absolute fact that the Nazis did kill thousands and thousands and thousands of innocent men, women and children, but not using nuclear weapons)

There is only one country with actual blood on their hands here with regard to nuclear weapons - the other nuclear powers have so far been able to show restraint.

As such, I find it pretty objectionable for people to suggest that it was "far better" this way, when the evidence really does not back it up.

I am not saying it is not best possible outcome for the world. Could it have been worse if the Nazis had for example nuked London in 1945 and killed a million? Sure, of course that might have happened, but it didn't actually happen. Perhaps had the nazis had that chance, the UK would have surrendered and there would have been peace and countless lives could have been saved and a completely new era of peace and prosperity begun? Or perhaps it would have also been untold slaughter and misery like the US did to Japan?

We just cant know, and so I object to people saying it was "far better" for history to have played out the way it did based mainly on - I suspect - the plot of Hollywood movies they've seen. History is written by the victors.

Anyway, this is way off topic and Godwin's Law has clearly been invoked. We should stop.

Post reply on HN