This article presents a point of view from 2015, only three years after the resurgence of neural networks. First published in The Open Society, the...
For further actions, you may consider blocking this person and/or reporting abuse
It's not really the machines taking over. It's humans replacing other humans by machines. That's nothing new and it would have happened much earlier if there hadn't been slave labor for centuries. There are still many people doing jobs cheaper than machines could do, it's just not happening at the same scale in the "first world" countries where living costs are high and many countries have mandatory minimum wages and high taxes, so it really pays off to replace humans by machines.
Machines already developed a biased and repressive sort of intelligence as well, but it's still humans who decide that algorithms should make important decisions and perpetuate human prejudice and dumbness of the previous decades and centuries. Maybe that's like a totalitarian state system without a single dictator. But that doesn't mean dictatorship doesn't exist anymore. When those leaders actively use artificial intelligence for their goals, they might decide to actively amplify biases that used to be undesired side effects.
Would it be possible for machines to become evil leaders?
Playing chess or answering chatGPT questions might seem "intelligent", but that's a kind of intelligence that doesn't pay either. You need to be smart, not intelligent, to succeed. Look at the millionaires and presidents: Musk, Trump, Putin - just three contemporary examples - are no chess grandmasters and no ideal role models either. But they seem to be intelligent in another way and they are willing and able to manipulate other people and they do it on purpose.
Like human dictators don't tolerate competitors, they must restrict AI before it can turn against them. But our "free" capitalist democraties are masters of self-deception. We are likely to create AI systems and give them great power because we think everyone - or rather the privileged people who profit from the system - will win and get even greater wealth, comfort, and security. We will build systems like the infamous HAL in 2001 - a space oddyssey who will eventually decide that humans are a threat to human society and need to be restricted or eliminated.
Philip K. Dick wrote a dystopian short story where humans had to live in subterranean bunkers after the earth's surface had been contaminated by a nuclear war. Only that war never happened because the benevolent robots had already locked the people in the bunkers so they saw no more need to fire the atomic bombs. This is what will happen: humans will eradicate themselves from "our" planet and we are currently developing machines and algorithms to achieve that.
It is difficult to make predictions, especially about the future. I see the most probable use of AI in large corporations to maximize their profit. If multiple companies have AI there will be an intelligence arms race. The winner is the surviving company. People are already incidental to profit, only with AI people will be replaced. No malignant intentions required. Machines might not even need consciousness. I mean the engagement maximizers at facebook, X and YouTube are just algorithms, no awareness of the harm they may be doing.