Pentagon's AI initiatives accelerate hard decisions on lethal autonomous weapons

https://apnews.com/article/us-military-ai-projects-0773b4937801e7a0573f44b57a9a5942

by haltist • 3 years ago

22 7 3 years ago

For anyone who has been paying attention, the obvious next step in drone warfare that is going to happen, is that Ukraine is going to develop ai terminal guidance for their drones to counter Russian point defense electronic warfare.

The Russians are putting a lot of jammers on vehicles which disables fpv drones when they get too close so they can’t land an rpg hit. The obvious counter to Russian EW is not needing a radio connection in the first place.

I would assume once they have terminal guidance figured out they will keep on trying to extend the range and ability of the system to detect its own targets so that Russian area denial ew is also rendered ineffective.

By the time they have that figured out, you have a lethal autonomous weapon.

So I’d say, anyone who is concerned about the ‘hard decisions’ of lethal ai is living in the past, we have the means, the technology and with the existential nature of the war in Ukraine, every incentive to build these weapons, so we will.

3seashells• 3 years ago

They have the video, they have the operator input, they have ground truth from observation drones for 2 years.thus There already is an AI trained on it?

T-A• 3 years ago

Russia already did that:

https://defencesecurityasia.com/en/lancet-russia-targetting/

atemerev• 3 years ago

We already have autonomous weapons: mines. The new ones will just come closer on their own. Like torpedoes already do.

I don’t understand what is the big deal, autonomous weapons are already used and obviously will be used later much more often. Like mines, they will be somewhat regulated; like with mines, regulations will often be ignored.

halJordan• 3 years ago

You do understand what the big deal is. If nine year old Mia dies on a ferry because the captain got too close to a mine then that's a mistake we could've prevented at several levels.

If Mia dies walking to school because not only is education necessary, but truancy illegal and the mine-equivalent decided she was in fact a vehicle theres nothing that really could've been done except not use the ai.

mdhb• 3 years ago

If anyone is actually interested in this topic in more depth this is maybe one of the best resources out there on it to help explain what the options look like and what trade offs are under consideration because a lot of the stuff people tend to refer to on this topic are pretty detached from reality https://youtu.be/1-0L5Wv86fQ

cyanydeez• 3 years ago

I just Assume they'll just "human-gap" decision mAking, like car companies pretend humans are a safe fall back when self driving runs into problems.

it'll still end up with AI making more and more complicated system and users trying to QAQC on the fly.

will not result in long term benefits and will most certainly be just a phase until the human-gap has aged out of the necessary real world training.

Pentagon's AI initiatives accelerate hard decisions on lethal autonomous weapons

Related Stories