Live data from Hacker News

Leaked Emails Show Google Expected Military Drone AI Work to Grow Exponentially

theintercept.com

421–430 of 527 posts

Re: Leaked Emails Show Google Expected Military Drone AI Work to Grow Exponentially

#421
post #309
post #140

Earlier quoted context omitted.

>Because they oppose the drone warfare program, which is killing noncombatants in nominally allied countries. I don't understand this logic. Drones are a weapon. They are not any more evil than any other weapon that the military uses. They aren't killing people autonomously. There is a human pilot behind every death. The pilot just happens to be hundreds or thousands miles away from the aircraft. Would we be fine wit…

"there is a human pilot behind every death." That statement won't last long term.

No, it'll become "there's a Google programmer behind every death".

Re: Leaked Emails Show Google Expected Military Drone AI Work to Grow Exponentially

#422
post #103

Earlier quoted context omitted.

> There is nothing that will stop this. The military will find a defense contractor to do their bidding. But Google can voluntarily elect to remove themselves from this sector. Let some other company do it. Let someone else figure it out. The blood of dead innocent civilians will be on their hands. You don't see Raytheon, Boeing, or Northrop crying about it. I work at Google but opinions are my own. This is actually…

Because they oppose the drone warfare program, which is killing noncombatants in nominally allied countries. https://m.huffpost.com/us/entry/us_561fafe2e4b028dd7ea6c4ff https://www.theatlantic.com/politics/archive/2016/03/the-oba... It’s not the military that is evil, it’s this particular program in the eyes of many. http://www.bbc.com/news/world-us-canada-24557333 Disclaimer: I am also fortunate to work at Google an…

I've often heard that civilian casualties would be higher without drones. First result on google seems to support it: http://www.slate.com/articles/news_and_politics/foreigners/2...

Re: Leaked Emails Show Google Expected Military Drone AI Work to Grow Exponentially

#423
post #46

The irony of this particular piece is that they are publishing emails expressing concern about how the media might frame and distort such an item and doing just that. The email discussion is mainly about potential revenue if they were to win the JEDI contract it's not about maven or 'expecting the drone ai work to grow'. Also worth noting that Google isn't the only contractor for maven yet they are made as such in mo…

Google probably can't change that, at least not overnight. The reason Google are experiencing leaks and taking flak for this is that it's only Google that ever claimed to be a superior moral actor. Amazon and Microsoft, for all their faults, never claimed to be much more than tech companies that want to make money - the extent of Microsoft's moral ambition was "computers are good, let's put a computer in every home",…

> to point out that men and women have different interests

I was nodding until this. Even if you agree with Damore, you can't pretend his argument was "men and women simply have different interests."

Re: Leaked Emails Show Google Expected Military Drone AI Work to Grow Exponentially

#424
post #103

Earlier quoted context omitted.

Because it is a conflict of interest. The public has put their trust into Google for 20 years. We supported them. We cheered them on. We have voluntarily given them all our personal data. Our emails. Our instant messages. Our voices. All of it feeds into their massive database for them to mine. We put our faith into them, in the hopes that they would honor their creed, that they would do no evil. But gradually, that…

> There is nothing that will stop this. The military will find a defense contractor to do their bidding. But Google can voluntarily elect to remove themselves from this sector. Let some other company do it. Let someone else figure it out. The blood of dead innocent civilians will be on their hands. You don't see Raytheon, Boeing, or Northrop crying about it. I work at Google but opinions are my own. This is actually…

Disclaimer: I also work at Google

I think that the problem is more specific than working with the military or developing weapons, and it's important to disentangle the more broad concerns with military work versus the concerns about this specific project occurring at Google. I personally would not like to work on developing weapons but it's definitely much harder to defend a position such as "building ships for the U.S Navy is evil".

The problem with Project Maven is plugging A.I. into drones. Sure, you can argue that A.I. today is limited and the technology isn't there to make them autonomous. However, perception/object detection is a huge component of what would be needed as part of an intelligence that can run a drone autonomously.

Then you arrive at basically 2 questions

1. Is Google's participation Project Maven making it easier for the Department of Defense to develop fully automated drones, or will they always leave a human in the loop?

2. Is enabling the Department of Defense to develop automated drones ethical?

#1

Yes, Project Maven enables automated drones. Sure, the Department of Defense could get automated perception technology from somewhere else, but that is in fact what the parent post is proposing Google should let the Department of Defense do. Also, A.I. advances very quickly and if the project continues there may be other ways in which Google is enabling full automation of drones, for example reinforcement-learning based control. I personally do not trust that the Department of Defense will not develop automated drones if they have the capability to do so. This is because they will probably see it as an arms race, and automating drones will serve most of the aims of their organization far too well to be put off indefinitely.

#2

I believe it is unethical to facilitate the development of fully autonomous, intelligent weapons. It's hard to make a stronger case than the video someone else has linked elsewhere in this thread: https://www.youtube.com/watch?v=TlO2gcs1YvM -The potential for chaos and destruction is very high. -The capability to find and eliminate targets will likely precede the ability to make ethical decisions, and human ethical oversight of the kill decisions will be reduced especially if the drones are scaled up For more on this check out http://autonomousweapons.org/

Re: Leaked Emails Show Google Expected Military Drone AI Work to Grow Exponentially

#425

I don't see anything wrong for a company of a particular country helping their military. GE, Siemens, Mitsubishi and all tech giants of that era did that with pride during WWII. Then why can't Google, Facebook for USA? Perhaps the illusion of globalization is making us perceive Google, Facebook and other international giants as global companies. Although, geniuses from around the world have significant contributions…

I think it's because of idealism innate in computing. Those young aspiring minds dream about making a positive change in the world, studying day and night, sacrificing pleasure for knowledge and capability, and at some point somebody will show them a video of a person blown into pieces and telling them: "Congratulations! This is the main fruit of your life. Well Done!" - moments later - "Why do you complain? You signed a contract so shut up!"

Re: Leaked Emails Show Google Expected Military Drone AI Work to Grow Exponentially

#426
post #397

Earlier quoted context omitted.

There is no challenger for "ultimate political power" within the US region How quaint - "regions" don't exist with modern weapons. To say the US doesn't have to worry in it's "region" is laughably naive. Their spending and investment is well beyond what is necessary to keep up with the "competition" Also quaint. China is able to leapfrog decades and trillions of dollars of US spending on defense due to modern technol…

While in general I agree with your realpolitik viewpoints this example was perhaps not the best thought out when rationalizing the application of US military resouces: "not to mention daily active engagements over Syria/Iraq/Afganistan etc." That's like saying it's ok to first smash windows and then charge for replacing them and being happy about how good you are in mending broken windows. It would be better no to br…

this example was perhaps not the best thought out when rationalizing the application of US military resources:

I agree :) However, my point was that these regional conflicts, with "tiny* countries are anything but. Superpowers are funding these wars, and it's all part of a larger game of hard and soft power. Saying the US's only worry are nukes, is just ... insane.

"While US is a stabilizing force in eastern Europe and the pacific, it has totally fucked up the lives of millions in the middle east area by it's interventionist escapades"

Excuse my vulgarity but shit has been fucked in the middle east for over 100 years. The US din't break any windows, if anything they bombed these broken windows into dust, along with everyone else.

In 10-20 years, this same statement will apply to the Pacific as China becomes more "assertive". To say the US "got it right" here is only because China's aggression or "assertiveness" is so new.

"It's as if they all believe that if you just give rebels guns then a George Washington will be magically summoned to lead the country in good order to a proper form of post-despotic administration."

I agree this is what is "sold" to the US public and the world. However the real motivations are always monetary, even in the case of the American Revolution.

Re: Leaked Emails Show Google Expected Military Drone AI Work to Grow Exponentially

#428
post #405

Earlier quoted context omitted.

Weapons are not all equal. Chemicals can be used as weapons. Most of society generally agrees not to participate in chemical warfare. Some weapons are more evil than others.

> Most of society generally agrees not to participate in chemical warfare. because of uncontrollable collateral damage. Same with mines that can't be deactivated. Drones, on the other hand, promises precise strike, with little to no collateral damage. If killing of "enemy combatants" can't be stopped, why not make it as precise and accurate as possible, and prevent collateral damage?

You do realize that drones are not precise at all? Plenty of collateral damage from them.

That is they are precise in targeting a specific location but they blow up everyone at that location?

Let's not even talk about double tap drone attacks. The second drone kills everyone attempting to help the wounded. Talk about do no evil...

Re: Leaked Emails Show Google Expected Military Drone AI Work to Grow Exponentially

#429

Earlier quoted context omitted.

When you are dying perhaps your definitions change. I hold the controversial (and often dismissed) opinion that Google is dying. I assert they are dying in part because they built a company around search advertising before they knew how to build a company, their initial success gave them the confidence and agency to create structural impediments to their ability to do anything but search advertising. And the search a…

Define dying ? I don't see how Google can go out of business in the next 10 years given their cash reserves. Eg: If they lost 2 billion dollars every year they could still keep doing that for decades. I'm pretty confident that they'll still be around in some form 20 years from now as well.

I define dying as being on a path toward non-existence. And I agree that given their cash it can be a very long very slow path. Sort of like Xerox or IBM or even Yahoo!. And just because there may be panic in the executive cubes because none of their strategies have worked so far, they may in fact find something to reboot them, wiser and stronger than before. Microsoft seems to have pulled it off and all it required was replacing the senior management and tossing out everything the company originally held sacred (like the Windows franchise). IBM has done it in the past but are struggling in this round.

I think the cash is an interesting challenge for its own reason though. Historically, public companies that held large fungible reserves like cash, were vulnerable to so called "Wall Street Raiders." These are people who would do the math and figure out that the assets of the company were worth more then the shares would indicate, they would buy up a controlling interest and then take apart the company, paying themselves a tidy profit. Traditionally that can't work on Google because the voting power of the class A stock is so much higher than the class B stock, however if the stock price were to crash, it becomes possible to make a play for all of that cash. The math is convoluted, and the founders need to continue their current steady sales of stock (which becomes class B when they sell it) but that flip happens before they run out of cash. I wonder sometimes if the IBM buyback of its own stock has been a plan to prevent this from happening there.

But timing the demise is secondary to the awareness on the part of the senior management that they can see their demise as a real and tangible thing. That awareness leads to stress and a willingness to do things counter to their established history in order to facilitate that survival and mitigating that threat to their existence.

The original question was "Did their moral compass shift?" and my observation is that the fear of death is a large metallic object which, when it gets too close to ones moral compass, deflects the needle in perhaps unexpected ways.

That Google's fear of dying is growing and their moral compass is being subverted in the name of survival is a hypothesis for a theory that explains their current behavior. I believe that the theory is both testable and falsifiable given information that we'll get going forward. As any good theory should be predictive, this theory predicts that as they fail to find a successful new business to supplant search advertising the ways in which they attempt to make money will become more and more disconnected from what their corporate culture claims to value. Further, if they do find a successful side business, their corporate cultural claims will re-emerge and they will once again make choices based on their cultural aspirations rather than for the sake of additional revenue.

If they continue to do more and more evil things, even after their future survival is no longer in doubt, it will show that my theory is completely wrong, they were simply evil to begin with and dying or not dying had nothing to do with the public shift in their willingness to be evil.

Re: Leaked Emails Show Google Expected Military Drone AI Work to Grow Exponentially

#430
post #397

Earlier quoted context omitted.

While in general I agree with your realpolitik viewpoints this example was perhaps not the best thought out when rationalizing the application of US military resouces: "not to mention daily active engagements over Syria/Iraq/Afganistan etc." That's like saying it's ok to first smash windows and then charge for replacing them and being happy about how good you are in mending broken windows. It would be better no to br…

this example was perhaps not the best thought out when rationalizing the application of US military resources: I agree :) However, my point was that these regional conflicts, with "tiny* countries are anything but. Superpowers are funding these wars, and it's all part of a larger game of hard and soft power. Saying the US's only worry are nukes, is just ... insane. "While US is a stabilizing force in eastern Europe a…

I agree with all you wrote above :)
Post reply on HN