Follow Slashdot stories on Twitter

 


Forgot your password?
Close
typodupeerror
×
The Military AI United States

The US Military's AI 'Swarm' Initiatives Speed Pace of Hard Decisions About Autonomous Weapons (apnews.com) 70

Posted by EditorDavid from the war-games dept.
AI employed by the U.S. military "has piloted pint-sized surveillance drones in special operations forces' missions and helped Ukraine in its war against Russia," reports the Associated Press.

But that's the beginning. AI also "tracks soldiers' fitness, predicts when Air Force planes need maintenance and helps keep tabs on rivals in space." Now, the Pentagon is intent on fielding multiple thousands of relatively inexpensive, expendable AI-enabled autonomous vehicles by 2026 to keep pace with China. The ambitious initiative — dubbed Replicator — seeks to "galvanize progress in the too-slow shift of U.S. military innovation to leverage platforms that are small, smart, cheap, and many," Deputy Secretary of Defense Kathleen Hicks said in August. While its funding is uncertain and details vague, Replicator is expected to accelerate hard decisions on what AI tech is mature and trustworthy enough to deploy — including on weaponized systems.'

There is little dispute among scientists, industry experts and Pentagon officials that the U.S. will within the next few years have fully autonomous lethal weapons. And though officials insist humans will always be in control, experts say advances in data-processing speed and machine-to-machine communications will inevitably relegate people to supervisory roles. That's especially true if, as expected, lethal weapons are deployed en masse in drone swarms. Many countries are working on them — and neither China, Russia, Iran, India or Pakistan have signed a U.S.-initiated pledge to use military AI responsibly.
This discussion has been archived. No new comments can be posted.

The US Military's AI 'Swarm' Initiatives Speed Pace of Hard Decisions About Autonomous Weapons More

The US Military's AI 'Swarm' Initiatives Speed Pace of Hard Decisions About Autonomous Weapons

Comments Filter:

  • AI written by the military would _never_ adopt a "kill them all and left god sort them out" attitude?

    Hopefully AI will take all the troublemakers out via rapture and leave the rest of us to a hundred years of peace...

    What could possibly go wrong?

    • war plan selection USA 1st strike!

    • From an essay I wrote on this in 2010: https://pdfernhout.net/recogni... [pdfernhout.net]
      "There is a fundamental mismatch between 21st century reality and 20th century security thinking. Those "security" agencies are using those tools of abundance, cooperation, and sharing mainly from a mindset of scarcity, competition, and secrecy. Given the power of 21st century technology as an amplifier (including as weapons of mass destruction), a scarcity-based approach to using such technology ultimately is just making us all insecur

      • I'm convinced John J. Mearsheimer's Offensive Realism theory is correct.
        You can't make predictions 10 years out with a wrong theory: "Why China Cannot Rise Peacefully" https://www.youtube.com/watch?... [youtube.com]
        Don't want to watch the video? TLDR it is basically the same argument in Batman vs. Superman
        Batman = USA who fears that China will become too powerful. Though they are currently not breaking any international rules, they could in the future.

    • Certainly you can find that attitude within the ranks of any military, but it doesn't always govern the high-level strategic decisions, or weapons development. Nations, and by extension their armies, can have goals other than mass slaughter.

      Take, for example, the US invasion of Afghanistan. The first part, large-scale military operation with the goal of regime change. They achieved it with 10,000 Taliban casualties and 2,000 civilian casualties. Yes, there were a few "kill them all" types who deliberately t

    • Just wait until the tech is transitioned to local LE.

  • The problem is that our enemies (Score:3)

    by Tablizer ( 95088 ) on Sunday November 26, 2023 @07:44PM (#64033431) Journal

    ...may skip safety to gain an advantage or to save a buck. Suppose they send a swarm after us, but our counter-swarm is remote controlled. They set off an EMP burst to screw up all radio, and bots are left to fight on their own. The more autonomous ones would have a big advantage.

    • move to def con 2!

    • If anything we've learned from history is that the US also skips on safety, any finger pointing to countries like Russia, China or Iran is just being a big hypocrite as the US is actually the one that bullies others into not developing nuclear weapons, but is still developing newer nuclear weapons and expanding its arsenal. talk about being a big f-ing hypocrite.

      • Re: (Score:2)

        by HiThere ( 15173 )

        That's not being a hypocrite, that's being self-centered. Just like all the countries that *do* decide to develop nuclear weapons anyway are self-centered. The world would be a lot safer if nobody had nuclear weapons. (OTOH, there would probably be lots more wars.) But it's really hard to stuff a genie back into the bottle.

    • Re: (Score:3)

      by AmiMoJo ( 196126 )

      We are seeing remote controlled drones used in Ukraine to great effect, and Russia seems to be able to do very little about them. The smaller ones are almost impossible to detect and shoot down, and very cheap to procure. The Ukrainians are literally buying them from AliExpress, strapping grenades to them, and sending them in through windows and doors.

      When they become autonomous, swarms of suicide drones will be deployed in a similar fashion.

      Well equipped militaries will simply flatten buildings like Israel

    • Re: (Score:2)

      by ceoyoyo ( 59147 )

      Ah yes. Our enemies might do this, therefore we must do it before they do!

      The warmonger's motto.

  • How is this not a solved problem? (Score:5, Informative)

    by MacMann ( 7518492 ) on Sunday November 26, 2023 @08:06PM (#64033475)

    We already have autonomous weapons, and we've been doing a pretty good job keeping them in check so far. It's not a perfect record but far from something that I'd think is a problem.

    Take the Phalanx CIWS system as an example. This system doesn't have an IFF system because anything moving fast enough towards a friendly asset to be a threat is fired upon. There's exceptions written in so friendly aircraft can approach safely, typically by following a path designated for a safe approach, a path kept guarded so an enemy can't use it to send in missiles or something.

    In the airspace over Ukraine it's just generally a "no fly zone" and anything airborne that can't be ruled out as a bird is shot down automatically. This has resulted in some "blue on blue" incidents, mostly on the Russian side, but that's a risk taken even without automated systems.

    The primary rule that protects friendly forces and noncombatants from automated killing machines is that we don't use them where the risks of "blue on blue" or "blue on green" events could happen. In that case a human is put in the decision loop. That's not foolproof because not every human will get it right, but it does mean that we have a person that is capable of more complex decision making than some rigid algorithm.

    What is making automated systems important is that weapons can move at much greater speeds than in the past, meaning if a person is in the loop they may not be able to process the threat quickly enough to respond. If they do respond in time then it may be because the human is using an overly simplistic decision tree on whether to fire or not. An example of that is telling a sailor on a ship to fire upon any radar contact that comes from shore. That's going to be effective in protecting against enemy forces on land, not put friendly aircraft coming from other ships at risk, but could mean someone fleeing the war in a Cessna could end up getting shot to pieces. Crazy things like this has happened.

    https://en.wikipedia.org/wiki/... [wikipedia.org]

    One of the more notable events occurred on Midway when the pilot of an RVNAF Cessna O-1 dropped a note on the deck of the carrier. The note read "Can you move these helicopter to the other side, I can land on your runway, I can fly 1 hour more, we have enough time to move. Please rescue me. Major Buang, Wife and 5 child." Midway's commanding officer, Captain L.C. Chambers ordered the flight deck crew to clear the landing area; in the process an estimated US$10 million worth of UH-1 Huey helicopters were pushed overboard into the South China Sea. Once the deck was clear Major Buang approached the deck, bounced once and then touched down and taxied to a halt with room to spare.[24] Major Buang became the first RVNAF fixed-wing pilot to ever land on a carrier. A second Cessna O-1 was also recovered by USS Midway that afternoon.[6]:âS121âS

  • The AP article lies that the US has used drones to attack Russia in the current Russo-Ukrainian war. The linked article makes no such claim except in the headline. And thank God for that. Editors, check your sources better and be way slower to post inflammatory headlines. This lie will make it halfway around the world, and will be believed by all the people whose belief will cause harm to pretty much everyone. Shame on you and shame on AP.

  • I'd rather have a military AI make a fire decision than a cop in the South.
    Those racist "thin blue line" bastards just want to kill minorities.
    Military AIs take out actual threats.

    The "thin blue line" is a way to make "us vs them" legtimate. It's not. Cops that live this way should be fired or shot by military AIs.
    If you're an "us vs them" person in law enforcement, just shoot yourself. Nobody will cry.

    • Community policing is better. Not much better, but better. But we're doing our best to reverse those gains and remove the educational requirement for cops so we can start hiring high school bullies again.

  • What if this gives birth to something like the Terminator, or something like Fred Saberhagin's Berserker series of sci/fi books???
    • I have been thinking, that it is possible, that humans are actually the toughest fighting machines ever. Yes, a terminator would be able to win over a single human, but think about how much other stuff humans can that terminator cannot. Your joints can heal on their own given enough time. No machine can do that. The will need a factory, a significant infrastructure for maintenance. Humans have millions of years of evolution behind them. Our bodies may not be particularly good at something specific, but they

  • So somebody involved is a Stargate fan and has a sense of humour.

    Because we're all looking to create an unstoppable swarm of tiny machines bent on destroying all biological life in the galaxy.

  • > There is little dispute among scientists, industry experts and Pentagon officials

    Scientists are people who follow the principles of science.
    Industry "experts" are just people employed who claim qualifications.
    Pentagon officials are dipshits with a starched shirt and some medals.

    NONE of them are qualified to opine on anything much. The so-called "scientists" might IF and ONLY IF that science field was related to what they opine about.

    AI is not a thing. There's no "intelligence" in AI. All there is is

    • Re: (Score:2)

      by HiThere ( 15173 )

      So, from your argument, you are not qualified to opine about intelligence.

      AI is a name. Names are always valid. A black cat can be named Grey, and that's still a valid name.

      The real problem with AI is that it's not a good clade. Very different things are given the same name. This is still valid English use, but it *is* confusing.

      As for "intelligence", there is no commonly accepted meaning. People just use the "I know it when I see it" test. (And if you mean IQ, there are AIs with a higher one than you

  • Land mines, anti-tank mines, and anti-shipping mines are fully autonomous and have been for over a century.
    Their "AI" is simple: if someone passes nearby, explode.
    Although modern anti-shipping mines are often designed to ignore decoys and take out high-value assets.

    • Land mines, anti-tank mines, and anti-shipping mines are fully autonomous and have been for over a century.
      Their "AI" is simple: if someone passes nearby, explode.
      Although modern anti-shipping mines are often designed to ignore decoys and take out high-value assets.

      This reminds me of a couple things, things that might not exactly follow where you were going.

      First thing this brings to mind is something of a joke, which may have some basis in reality. An admiral is in his office on the flagship of a navy flotilla when the relative calm is broken by a large explosion. He gets up to run to the bridge to see what has happened. A mine had exploded off to the side of the ship but did only superficial damage to the thick hull, even so this was a considerable danger since h

  • This is how you get an AI overlord. Naming it fucking "Replicator"... the fuck.

  • If there were ethics in war... (Score:3)

    by VeryFluffyBunny ( 5037285 ) on Monday November 27, 2023 @05:48AM (#64034277)
    ...we wouldn't have nuclear weapons. Their purpose is as a threat to inflict massive civilian casualties which is against the Geneva conventions. Get real; autonomous lethal robots used indiscriminately during warfare are a thing regardless of what our militaries & politicians claim. Much in the same way that "smart bombs" & state of the art targeting systems are used to target civilians. We only need to look at Yemen & Gaza to see that in action.
  • As a lot of people are calling for a pause on AI development, the sector where AI development is the most dangerous is continuing in developing AI. It's actually defense AI that is the most potential to end up in destroying us, not the AI used in civilian applications. Defense development will never pause AI development, so why should civilians have to, and it's actually defense AI that will be exempted from any regulation, mark my words.
  • John Conner is in danger..

  • OpenAI: Behold.. ChatGPT!
    People: OMG.. autocomplete will destroy civilization
    Pundits: ChatGPT is an existential threat to humanity
    People: F-35 F*** yeah!
    People: Oh.. .no.. AGI... oh no .. LLM.. oh no
    People: F-22 shoot them out of the sky! ...
    Meanwhile in the military: lets create swarms of AI robots to destroy civilization
    Military laughing at us: haha.. they're scared of autocomplete, but don't even blink an eye on the literal civilization ending weapons we routinely create. They even cheer us on! Haha.

Slashdot Top Deals

To write good code is a worthy challenge, and a source of civilized delight. -- stolen and paraphrased from William Safire

Close