Pentagon Pushes for Ethical AI (Updated 03/14/2020) Possibly Final Update Until More Information is Given

[FakeCIA]

https://www.military.com/daily-news/2020/02/24/if-its-not-ethical-they-wont-field-it-pentagon-release-new-ai-guidelines.html?utm_medium=Social&utm_source=Twitter#Echobox=1582596662

 

Air Force Lt General Shanahan, Director of the JAIC (Department of Defense) has issued a statement regarding the United States use of artificial intelligence in warfare. Shanahan said in a statement that "We will not field an algorithm until we are convinced it meets our level of performance and our standard, and if we don't believe it can be used in a safe and ethical manner, we won't field it." In other statements to reporters, Shanahan made comments that the use of AI by Russia and China raise serious concerns about human rights, ethics, and international relations. He also made comments that he does not want AI to be used by the military to track citizens as is practice in China. The Pentagon has pushed for more mandates regarding AI in recent years, but concerns that those guidelines won't be recognized internationally are there as well. Shanahan also made references to a speech made by Russian President Putin in 2017, saying "whoever becomes the leader in this sphere will become the ruler of the world."

 

 

My opinion: Contrary to popular belief, there are rules to warfare. For example, you cannot target any medical facility or transport. I think the move to a more regulated atmosphere is a good thing. AI already brings up a lot of ethics questions, but it is much more hotly disputed in terms of how it could be used in war. I believe that AI should be used in war to lower the amount of casualties, but it definitely needs to be regulated by an objective human entity and undergo regular ethics testing. SkyNet, Hal 9000, and Ultron are good examples of what could go wrong in the movies, but I think with the right infrastructure in place, there could be a better system than just blowing each other up. Any AI that is designed and is functional should be under heavy watch and have a kill switch. I think that this is a good move and Shanahan is probably the best guy to help pave the way. He's made comments before on this and is one of the advisors for this kind of subject to the CIA and White House. Obviously, we are not to the point yet that we have an AI capable of thinking on it's own completely, but it is good to think ahead.

 

 

Edited February 25, 2020 by FakeCIA
Added Video from Outside Source/Reporter in Briefing

[FakeCIA]

https://www.defense.gov/

 

To watch the original briefing and see the transcript, go to the official Department of Defense website here and click on the "What's New?" section. A transcript translated into multiple languages will be available soon on the site.

[FakeCIA]

1 hour ago, huilun02 said:

Probably the same people who wanted Google AI to be used to aid drone strikes in faraway places

There are some that want that, yes, as a deterrent. But, Google does not have authorization or access to advanced weapons systems. Google, Microsoft, and other companies are not permitted to make weapons themselves and must comply with US law in regards to weapons manufacturing. That is the reason JAIC exists, to oversee the ethical operations of AI in the United States and its assets. While Microsoft may make the computers for the aircraft carrier, they are not allowed to make the missiles.

[Doobeedoo]

There's more to it to regulate AI in a way as many companies agree. Though government and many humans are not ethical them selfs sooo...

[Donut417]

3 hours ago, FakeCIA said:

Contrary to popular belief, there are rules to warfare

Yeah it’s called the Geneva convention. 

[FakeCIA]

Update: Translated transcripts from the press briefing are now available in various languages on the official Department of Defense site based on region. Military.com, the VOA, and the Freedom of the Press Association also have transcripts available. If your language is not shown as an option, third-party extensions like Google Translate are available for all of these sites and are fully supported. You may also request a transcript directly from the DoD via their contact page.

[FakeCIA]

19 hours ago, Doobeedoo said:

There's more to it to regulate AI in a way as many companies agree. Though government and many humans are not ethical them selfs sooo...

I think that just the fact the Pentagon even had a press briefing about this is amazing. It shows that they are working on public trust both international and domestic. They are known for being very secretive, especially since 9/11. In the press briefing, they were very transparent with how this will go forward and what they plan to do, more so than I've ever seen them. We still don't know many of the details, but this seems like they are trying to gain the public's trust for a new way of ethical warfare and other applications. I think they are just as unsure as we are and want to approach this with the backing of the public.

[hobobobo]

I wonder if they themselves belive all that bullshit about ethics, human rights etc. I mean, talk about hypocrisy

[FakeCIA]

2 hours ago, hobobobo said:

I wonder if they themselves belive all that bullshit about ethics, human rights etc. I mean, talk about hypocrisy

We will have to see. Meanwhile, I'll start getting my bunker ready. I do understand your concern though. I'd be lying if I said I was perfectly fine with the Pentagon getting involved with this subject directly, considering the organizations responsible for MK-Ultra and Agent Orange hold power in this, even if it is heavily restricted. I want a board of objective individuals to oversee all of this from an outside perspective. I'd be comfortable with the United Nations making a committee similar to JAIC to oversee all AI operations in military use. There could be a separate committee for public use AI.

[FakeCIA]

The Age of AI S1-E1 from YouTube Originals hosted by Tony Stark, no wait... Sherlock Holmes, crap no, Robert Downey Jr.

[hobobobo]

46 minutes ago, FakeCIA said:

We will have to see. Meanwhile, I'll start getting my bunker ready. I do understand your concern though. I'd be lying if I said I was perfectly fine with the Pentagon getting involved with this subject directly, considering the organizations responsible for MK-Ultra and Agent Orange hold power in this, even if it is heavily restricted. I want a board of objective individuals to oversee all of this from an outside perspective. I'd be comfortable with the United Nations making a committee similar to JAIC to oversee all AI operations in military use. There could be a separate committee for public use AI.

The thing is, there is no way to control the way AI is going to be implemented into warfare. All the surface-level stuff with pentagon doesnt really matter, what concerns me is special branches of DAPRA working on advanced weaponry, they literally have 0 oversight or ethical restrictions until the reasearch is turned into a tangiable weapon. Same goes for my country, kinda terrifies me a bit, considering that russian military-industrial complex manages to produce about 30-50% of output of american counterparts on 0,1x budget. Cant wait for drone swarms with ai targeting and threat assesment

[FakeCIA]

1 hour ago, hobobobo said:

The thing is, there is no way to control the way AI is going to be implemented into warfare. All the surface-level stuff with pentagon doesnt really matter, what concerns me is special branches of DAPRA working on advanced weaponry, they literally have 0 oversight or ethical restrictions until the reasearch is turned into a tangiable weapon. Same goes for my country, kinda terrifies me a bit, considering that russian military-industrial complex manages to produce about 30-50% of output of american counterparts on 0,1x budget. Cant wait for drone swarms with ai targeting and threat assesment

You're invited to my bunker.

[Sauron]

On 2/25/2020 at 12:44 PM, FakeCIA said:

if we don't believe it can be used in a safe and ethical manner, we won't field it.

Lmao, they never cared about this. They're just butthurt because China got it first.

On 2/25/2020 at 12:44 PM, FakeCIA said:

SkyNet, Hal 9000, and Ultron are good examples of what could go wrong in the movies

No they're not, our "AI" is nowhere near being sentient. The risk here is that it will display unforeseen behavior because of insufficient or incorrect training. Though to be honest even the idea of using it for military purposes sickens me, even if it worked perfectly.

[FakeCIA]

1 hour ago, Sauron said:

_snip_

Technically, China is using a facial recognition program that details specific signatures in one's facial features. It isn't really considered an AI by many people since it can't react to those features. And in my first comment on the first post, I did say we are nowhere near sentient, but we are headed in that direction. Movies like Avengers: Age of Ultron show how our fear of not understanding can push us to create flawed systems if we aren't careful. I see an artificial intelligence as something like Mother from the Netflix movie, I Am Mother. In that film, after the destruction of humanity, an AI is tasked with raising a human, from infancy to adulthood, exploring the complexities of how that could be accomplished.

[Doobeedoo]

4 hours ago, FakeCIA said:

I think that just the fact the Pentagon even had a press briefing about this is amazing. It shows that they are working on public trust both international and domestic. They are known for being very secretive, especially since 9/11. In the press briefing, they were very transparent with how this will go forward and what they plan to do, more so than I've ever seen them. We still don't know many of the details, but this seems like they are trying to gain the public's trust for a new way of ethical warfare and other applications. I think they are just as unsure as we are and want to approach this with the backing of the public.

I mean yeah makes sense too being how AI will become so ubiquitous and widespread eventually. 

[Sauron]

3 minutes ago, FakeCIA said:

China is using a facial recognition program that details specific signatures in one's facial features. It isn't really considered an AI by many people since it can't react to those features.

What do you mean by "react"? Right now recognizing faces is more or less the best you can do with AI.

7 minutes ago, FakeCIA said:

I did say we are nowhere near sentient, but we are headed in that direction.

No, we are definitely not. We'll get faster and more accurate AI but we'll never get anything resembling sentience unless there's some drastic and unforeseeable change in the technology. This is like saying that if you keep developing car technology eventually you'll get a spaceship.

11 minutes ago, FakeCIA said:

Movies like Avengers: Age of Ultron show how our fear of not understanding can push us to create flawed systems if we aren't careful. I see an artificial intelligence as something like Mother from the Netflix movie, I Am Mother. In that film, after the destruction of humanity, an AI is tasked with raising a human, from infancy to adulthood, exploring the complexities of how that could be accomplished.

I can see why you'd be confused if you get all your knowledge of AI technology from fictional movies. Do you also think robots follow Asimov's 3 laws of robotics?

[FakeCIA]

1 hour ago, Sauron said:

snip

I use fictional movies because that is the general public's best understanding of where the technology will go. I read reports from MIT, Boston Dynamics, and other institutions too, but I like seeing a working example of what other people are thinking. I'm not just watching these movies for shits and giggles, I'm watching them because they are conceptual examples of the future. As for China's facial recognition program, it can't react like you or I to stimuli. That program can see a certain facial feature and mark it in a way it has been programmed to do. It can't mark it, then go and have a different reaction than those set for it. It can't choose what it wants to do with that information.

[FakeCIA]

https://www.defense.gov/Explore/News/Article/Article/2094085/dod-adopts-5-principles-of-artificial-intelligence-ethics/

 

Update: The DoD outlines the 5 laws of Artificial Intelligence Ethics

[Sauron]

15 minutes ago, FakeCIA said:

I use fictional movies because that is the general public's best understanding of where the technology will go.

I don't care about what the general public understands, I care about what the technology can actually do.

17 minutes ago, FakeCIA said:

I'm not just watching these movies for shits and giggles, I'm watching them because they are conceptual examples of the future.

That doesn't mean anything. The people who wrote Avengers aren't computer scientists and even if they were they wouldn't have cared about being scientifically accurate - the point wasn't to predict the future, it was to write a good story. Not to mention characters like Ultron or Skynet were originally conceived loooong before current AI technology surfaced. "What if computer sentient???" has been a staple of science fiction writing for decades, pretty much since the invention of computers, that doesn't mean it's a correct assessment of what the future holds.

23 minutes ago, FakeCIA said:

As for China's facial recognition program, it can't react like you or I to stimuli.

That would be because it's not sentient.

24 minutes ago, FakeCIA said:

That program can see a certain facial feature and mark it in a way it has been programmed to do. It can't mark it, then go and have a different reaction than those set for it. It can't choose what it wants to do with that information.

Yes, that is what current AI technology is. We don't have an artificial general intelligence that can handle things it wasn't designed to. If there are only a handful of possible decisions you can make once the face has been recognized then training the AI to choose what to do is very inefficient when you could just hardcode it.

[FakeCIA]

1 hour ago, Sauron said:

 

You literally just read all the points I made back at me. Are you reading this thread and the above comments? The technology is going to be used by and on the general public so it is their understanding that matters most, not scientists or military leaders, those are secondary. The writers of the Avengers movies are not scientists, and that is exactly my point of who the technology is for and what the idea is for. The technology is going to become sentient, whether it is tomorrow or in 1,000 years. I can't predict when, but I can see it happening based on the current state of ideals and direction people want to take. China can say it has an AI just to flex, that is entirely up to them, but it's not correct, not when we have so far to go still.

[Sauron]

1 minute ago, FakeCIA said:

You literally just read all the points I made back at me.

No I didn't and how you came to that conclusion baffles me.

2 minutes ago, FakeCIA said:

Are you reading this thread and the above comments? The technology is going to be used by and on the general public so it is their understanding that matters most, not scientists or military leaders, those are secondary.

That's bullshit, regulations should exist to increase safety, not to appease people who have no clue what they're talking about. If you want to learn how AI actually works you can open a wikipedia page.

4 minutes ago, FakeCIA said:

The technology is going to become sentient, whether it is tomorrow or in 1,000 years. I can't predict when, but I can see it happening based on the current state of ideals and direction people want to take.

So your argument is that because some people who know nothing about AI think it may get sentient one day, then it will get sentient? I'm sorry but this is some of the most ridiculous nonsense I've ever read.

5 minutes ago, FakeCIA said:

China can say it has an AI just to flex, that is entirely up to them, but it's not correct, not when we have so far to go still.

We've had "AI" since the '60s, if a machine is capable of looking at a set of data and making a decision then it qualifies. What you're thinking of is Artificial General Intelligence, which isn't what anyone is talking about here nor something that will automatically evolve out of existing AI research. When the US army says it wants to use AI they mean they want to use facial recognition to spot drone strike targets.

[FakeCIA]

1 hour ago, Sauron said:

-snippity snap-

I don't think you quite understand the points I'm trying to make, that's ok, and I don't think you have read the articles or have watched the videos posted given how fast you are responding. Anyway, I'm keeping an eye on this thread for updates about this situation and further news and will update accordingly. You can post your opinion about the facts all you want, just don't start ranting and raving like Alex Jones or mods will move this to "Off Topic," which I don't want.

[Tristerin]

On 2/25/2020 at 5:44 AM, FakeCIA said:

 

My opinion: Contrary to popular belief, there are rules to warfare. For example, you cannot target any medical facility or transport. I think the move to a more regulated atmosphere is a good thing. 

 

 

To this statement I laugh out loud.  So "we" slap ourselves on the wrist when we bomb hospitals?  Nope not even then.

 

Does no one remember, in such a short time ago, the Kunduz Hospital Airstrike?  I can go into the details but we police ourselves and a short google search away you will see nothing came of this.  A report was made, after burning people alive in hospital beds.  

 

Anyhow I digress - The words ETHICAL and GOVERNMENT do not mix.  What they say, and what they will actually do are completely different things.

[FakeCIA]

On 2/25/2020 at 7:37 AM, Donut417 said:

Yeah it’s called the Geneva convention. 

There are official rules to warfare, especially after World War 2. The Kunduz airstrike was due to inaccurate intel and confusion between Taliban and Afghan forces, and is a one-time isolated incident that doesn't really fit as an example in this thread, while I do respect that example and thank you for bringing that up.

1 hour ago, Tristerin said:

To this statement I laugh out loud.  So "we" slap ourselves on the wrist when we bomb hospitals?  Nope not even then.

 

Does no one remember, in such a short time ago, the Kunduz Hospital Airstrike?  I can go into the details but we police ourselves and a short google search away you will see nothing came of this.  A report was made, after burning people alive in hospital beds.  

 

Anyhow I digress - The words ETHICAL and GOVERNMENT do not mix.  What they say, and what they will actually do are completely different things.

 

[Sauron]

29 minutes ago, FakeCIA said:

I don't think you quite understand the points I'm trying to make, that's ok, and I don't think you have read the articles or have watched the videos posted given how fast you are responding. Anyway, I'm keeping an eye on this thread for updates about this situation and further news and will update accordingly. You can post your opinion about the facts all you want, just don't start ranting and raving like Alex Jones or mods will move this to "Off Topic," which I don't want.

The only one acting like Alex Jones here is you with the blatant disregard for scientific fact in favor of science fiction and the prepper talk...

3 hours ago, FakeCIA said:

Meanwhile, I'll start getting my bunker ready.

I mean, a little self awareness has never hurt anyone.