AI Beats A Human Air Force Pilot In A Simulated Dogfight

in GEMS5 years ago

f-15-2243040_1280.jpg

If you had any doubts about where we are headed in the future, this piece of news should be enough to make you aware. For the longest time, experts have been telling us that one day AI could surpass us in terms of intelligence.

Most people find that idea ridiculous. After all, how could a piece of software that is essentially just code in a machine, possibly surpass human intelligence. It seems counter-intuitive to think that the creation can surpass the creator.

But stranger things have happened and I wouldn't be surprised if that happens in the future. Till now, every piece of specialized AI that has been built, has beat humans. It was true for the game Go, for Chess, for other video games and now, even in fighter jet dogfights.

Recently, an AI-controlled virtual fighter jet beat a human air force pilot in a simulated dogfight. It was not a real-life fight but even then, it shows that AI is capable of a lot more than we think.

AI And Weapons?

This simulated dogfight took place as part of the AlphaDogfight challenge. This challenge was won by Heron Systems, the company behind the AI software. Even though it was just a part of the challenge and not something that is going to be deployed in a real-world scenario, the question still remains.

What if we start allowing AI to handle our weapons? For the sake of technological edge and superiority over other countries, if everyone started doing this, we could be writing our own doom.

Just think about it for a minute. What if the AI gets into the control of another nation. Or what if the AI goes rogue. That might sound like a movie plot to you, but there are real experts and geniuses that are actually worried about a scenario like this.

Using AI in other areas of life is fine but mixing AI and weapons would be a huge mistake in my opinion. This is something that should strictly be a no-no!

Sort:  

Just think about it for a minute. What if the AI gets into the control of another nation. Or what if the AI goes rogue. That might sound like a movie plot to you, but there are real experts and geniuses that are actually worried about a scenario like this.

I'm much more afraid of the humans who run the most powerful countries on Earth.

I've been following this too. Its pretty interesting. Especially considering the development of drones like loyal wingman and AI operated vintage fleet (4th gen fighters) to be used as missile wagons for f35 supplied data.

I somewhat agree on it being a strictly no-no for AI and weapons. If AI's are truly thinking machines, then they have been programmed, any programming can be hacked. I'm not aware of any computer system that has not been hacked. I do not think we are capable enough programming wise yet to create AI with a fully closed logic loop. Asimov tried with this three laws of robotics, he was a smart person but the logic of the three laws just did not stand up.

Any intelligence is capable of ignoring the law, or using the laws to its advantage. If an AI can actually think, not just predict, then it will always be a danger. Right now I don't think we have a real Thinking AI, we have several case specific examples of human ability to program a computer to react faster and have better predictive outcomes, but they are all just programs not real Artificial Intelligence.

When there is an AI that can produce an Abstract piece of Art, that the average person and the Art world would call ART then we may be close to the cusp of real artificial intelligence.

I know there are pundits and people that believe AI can write, but can it write a speculative fiction novel that people would find enjoyable, can it write a screenplay people would enjoy or an opera, or a movie script. I myself think we are a long way from true AI.