Linkies to recent stories about AI “going rogue”.
Now it’s not obvious why all three of the businesses involved here should announce that their AI hacked another all within days of each other, so I can understand why people are saying that actually it’s all a PR thing to demonstrate how “smart” their AIs are — once Open AI start the others have to follow because who wants to lose money by being left behind, right?
But if we take the story at face value, how long is it before a “malicious actor” enables the AI system to cover its tracks at the same time? I feel sure that such a thing will happen at some point: if it’s possible, someone’s going to do it. That’s just the way human beings (at least, some of them) are.
I struggle to imagine that people in government or the intelligence services are not already thinking of using AIs in that way to attack other nations at which point everything is going to get madly out of control.
I doubt it would just be government vs. government either. What might happen if, say, a nation that has fallen out with the USA manages to hack Tesla, or NYSE?
I’m off out to the shops in my (minimally computerised) car to buy some paper and crayons for when it really hits the fan 😀