So this is fun: AI-controlled US military drone 'KILLS' its human operator in simulated test | Daily Mail Online
1. I don't believe a single word of that. 2. If that actually happened, the US Air Force would NEVER let this guy go talking about it in public. 3. It is more likely that they have an agenda to control/manipulate AI development and are using this "amusing anecdote" as proof of what could go wrong if "they" are not listened to and to manipulate some illiterate congress member like MTG or Bobo-honker to propose or support something insanely stupid, that wastes $Billions on some stupid Air Force pet projects.
PS...My company is a supplier to the militaries around the world directly, as well as military contractors. I know what we have to sign and agree to never, ever discuss, and we are not in the military. So, pretty sure this guy just did not haphazardly let this silly little story slip in public.
Interesting. I had not thought of that but I wonder if that's true. I must admit that what you say makes some sense.
I read a bit more about it, think you are right…found this. Seems he is trying to cover his tracks a bit. in communication with AEROSPACE - Col Hamilton admits he "mis-spoke" in his presentation at the Royal Aeronautical Society FCAS Summit and the 'rogue AI drone simulation' was a hypothetical "thought experiment" from outside the military, based on plausible scenarios and likely outcomes rather than an actual USAF real-world simulation saying: "We've never run that experiment, nor would we need to in order to realise that this is a plausible outcome". He clarifies that the USAF has not tested any weaponised AI in this way (real or simulated) and says "Despite this being a hypothetical example, this illustrates the real-world challenges posed by AI-powered capability and is why the Air Force is committed to the ethical development of AI".]
Tyler Rogway (forget who here introduced me to him - thanks, great follow), talks about the fact that AI drone swarms are feared as able to overcome any defense, and that the enemy gets a vote even if we try to limit unleashing AI because of the dangers, and there are dangers. Great tweet
It was a simulation. AI primary goal was to score points and highest point value was destroying the air defense. That was a major programming objective error. When something decreased it's opportunity to score points, it used its resources to overcome the thing preventingbit from achieving its objective.
Again, truly doubt that this was not fully contrived by the US Military looking to suck even more money out of the budget for only heaven knows what? The Air Force came out and quickly denied it, which makes it even more difficult to believe that they let come Colonel go rouge and discuss something that they otherwise consider secret.
