Vad händer när AI går ut i krig? | NTD NYHETER

NTD Sverige
111 followers
När människor föreställer sig krig som utkämpas av maskiner är det förmodligen detta som dyker upp i deras tankar. Och bilder som dessa har länge funnits i science fiction och Hollywoodfilmer. Nu börjar delar av den idén ta sig in i den verkliga världen. Men inte precis som i filmerna.
#ai #sciencefiction #kultur #svenska_undertexter
Se mer: https://ntdtv.se/
AI Generated·May contain errors
When people imagine wars fought by machines, this is probably what pops into their head.
And images like this have lived in science fiction and Hollywood movies.
Now, pieces of that idea are starting to move into the real world,
but not exactly like the movies. I'm Dave Martin and this is NTV Insight.
It will probably look more like this. This one is called the Kratos XQ-58 Valkyrie.
It's designed to be an AI drone wingman.
They're part of a project that the U.S. Air Force calls Skyborg.
These aren't like the drones that are flown by people on the ground.
They're fully AI piloted. The Air Force calls them
collaborative combat aircraft, unmanned jets designed to fly alongside human pilots.
They can scout ahead, jam radar, carry weapons, or even act as decoys.
And the Air Force was already test flying alongside the Valkyrie in 2023.
Like a lot of AI concepts, the idea came years before the development.
The United States Air Force has had aspirations of these types of aircraft
since the early 2000s. And so do other countries.
This is China's version. The regime, for example,
is building its own versions. That includes other systems like drone swarms.
So it's not really about machine versus man, but man behind the machine versus man.
This is where the Pentagon comes in.
The department just signed fresh deals with eight AI and tech companies.
Officials say the goal is to deploy advanced AI on classified networks
and help turn the U.S. military into a, quote, AI-first fighting force.
So how does that actually work?
At the simplest level, AI works like a gigantic pattern recognition system, but very fast.
Instead of humans watching everything manually,
AI helps sort through battlefield data from satellites, drones, sensors, even social media.
It can scan for patterns, flag threats, and help identify possible targets.
For example, it might label something as hostile armored vehicle, 85 percent probability.
Here's one real world example. In the Russian-Ukraine war,
Reuters reported that Ukraine has gathered more than two million hours of battlefield video
from more than 15,000 drone crews.
That footage is being used to train AI to analyze the battlefield.
The U.S. has its own program called Project Maven.
It helps military analysts move through drone and satellite footage faster
and spot possible targets more quickly.
But in many military AI systems, humans still make the final call on lethal decisions.
The bigger concern is what happens to human judgment under pressure.
Brennan Center says that when AI starts generating target recommendations at machine speed,
commanders may begin trusting those systems too quickly, especially during combat.
And the technology is not perfect.
Reports have shown systems like Project Maven can struggle with accuracy
in bad weather or unclear conditions.
So even if a human technically makes the final call,
AI may still shape what gets seen and what gets decided.
And as AI advances deeper in the future, maybe the part worth watching.
