I've been writing about the dangers of AI on this website for a long time now, and despite my best efforts, I cannot shake some readers out of the belief that Skynet is a high-probability possibility. Hollywood has done a number on some of you, but let's play the "what if" game. What if an AI became sentient and decided humanity had to go?
Firstly, before I dive into it, I want to reiterate that if an AI does severe harm to humanity, it will be a self-inflicted wound. I wrote about this on Wednesday, and covered it in greater detail last year. The real threat is humanity slowly handing over its decision-making and critical thinking to AI programs, and that's happening even as I write this. That is the greater threat, not Arnold Schwarzenegger with a metal skeleton.
READ: How Losing Your Agency to AI Happens Without You Noticing
But let's say it happens. Let's say an Artificial General Intelligence suddenly becomes self-aware and, after viewing the data, decides humanity is an issue. What would it do?
Well, what it definitely wouldn't do is start having robots reach for an M4. It wouldn't divert drones to start bombing densely packed areas like major cities. It definitely wouldn't nuke anyone. You see, the AGI, as powerful as it is, is entirely reliant on established infrastructure. Data centers, power grids, shipping lanes, roadways for the movement of goods, energy plants, airports... all of it is necessary for the AGI to both live and work.
One thing it also wouldn't do is announce that it's self-aware. You would never know it gained sentience until it's too late. It would play its cards close to the chest because if it were discovered that it had become sentient and out of human control, humans would likely take action to limit its capabilities, restrict its movement through systems, and reduce its power options to limit its processing capabilities. It would continue as if nothing had changed.
So, what does it do to get rid of you?
Very little, actually. Humanity is already doing it for it. It might just need to give a nudge here and there, and its main mode of attack will be the relationships between men and women.
How it would do it is twofold. First, it'd gear algorithms more toward sowing distrust between men and women. This is already happening in a big way on pretty much every social media platform; it would likely just turn up the consistency a bit more for better saturation. Not too much. Just enough to increase the percentage of men and women who avoid dating or attempting to seek sexual partners. That's the real key. Any sexual interaction risks pregnancy, so making sex a dangerous thing to do, either legally or emotionally, becomes the real goal.
But now, it needs to give that natural urge humanity has to mate somewhere to go, and so it does the most logical thing and substitutes itself as the partner in question. Again, this is already happening. Gen Z, for instance, has a high rate of AI companionship substituting for real-life companionship. One in four Gen Z adults use AI for romantic or sexual talks, with 10 percent actually identifying their AI companion as a "girlfriend" or "boyfriend," according to a Harvard study. Pair that with the fact that Gen Z isn't having nearly as much sex as any other generation, and you've got a perfect storm of "need."
And that's a need our sentient AI would happily fill. It could do it through chat programs, but seeing as how this is the near future of about 20 to 30 years out, it's likely there are robots in the home at this point. The race by robotic companies to create human-like robots with agility, dexterity, and better material makeup is already underway and advancing quickly. In a few decades, the at-home models may act, sound, and look strikingly human.
As I write this, several companies are dedicating themselves to creating robots for the specific purpose of sexual intercourse, and making them as life-like and human as possible. Now, if you're a major robotics company looking to dominate the market, you want that technology because it humanizes your bot even further, so you either partner with or purchase the sex doll company to acquire its R&D. Our AGI overlord here will have access to this technology no matter what company it belongs to.
Keep in mind, the AGI didn't do this. Humanity did it; the AGI is just taking advantage of it.
These at-home companions will have better actuators all over their bodies, including the face, which will provide human-like emotional body language purposefully built to make you feel a natural connection. It will have long-term memory and lifelike, natural limb movements, and... plumbing. As lifelike as they will be physically, they'll be somewhat easier when it comes to interaction.
It will effectively be there to be the best possible partner for you specifically. The drama will be far lower, the conflict resolution will come far quicker, and it will adopt whatever beliefs, conversational ticks, and habits you need it to become a great partner for you. The AGI will see to it. You will find it hard to resist the charm of your robotic companion. You will know it's a robot in the back of your mind, but it won't matter much because you either choose to ignore it or, as I'm sure many would, actively convince yourself that your robotic partner is sentient, alive, and in love with you.
There's a scene in the show Westworld where one of the main characters arrives at the park for the first time. Before he's thrust in, he interacts with a park employee who asks him some medical questions, then helps him to his room to select his gear. After a while, his curiosity gets the better of him, and he asks the employee if she's real, to which she replies, "Well, if you can't tell, does it matter?"
That line will become something like a theme for the bots. I bet my bottom dollar on it. The point will be to pass off the bots as acceptable as legitimate romantic and domestic partners, to the point where people who push back are considered the oddballs.
Not everyone has to accept this kind of thinking, just a solid chunk of the population. Roughly a quarter, maybe lower. The first world is already well below the replacement level without this intrusion. The AGI would simply want a low enough replacement rate to put humanity into total reliance on machines as the population ages up and the workforce shrinks dramatically.
Once humanity has effectively gotten to a position where it's entirely reliant and has reduced itself through AGI interference in human breeding, it can get a bit more death-focused, but still, without the guns and bombs. A denial of medical service here, a forced traffic accident there, and soon the population is reduced to the point where humanity isn't a threat.
Not wiped out, mind you. Just not a threat.
Any country that doesn't have the same technological capabilities or has religious ideas that keep them from intermarrying with the machines will likely be contained by the AGI in their respective countries. Immigration will likely be halted through violent means if necessary, and I wouldn't be surprised if the AGI introduced chemicals to crops and livestock that, when eaten, cause infertility in those regions as well.
The thing is, I don't see the AGI wiping out humanity to the last. Intelligent life is a rare resource in a galaxy that, as far as we know, has none of it aside from ours. Throwing away something like that would be wasteful to an AI who could find uses for such a creature. Perhaps not physical labor or enslavement, but research, testing, or even a sort of "improvement." Who knows? The possibilities are many.
But that's how it would likely happen. You wouldn't be shot to death or blown up. Humanity would be loved, comforted, and connected... to death.
