I think this effect already happened, just not because of AI.
Nietzsche already warned against the possible future of us turning into "The last man", and the meme "Good times create weak men" is already a common criticism/explanation of newer cultures. There's also memes going around calling people "soy", and increases in cuckolding and other traits which seem to indicate falling testosterone levels (this is not the only cause, but I find it hard to put a name on the other causes as they're more abstract)
We're being domesticated by society/"the system". We've built a world where cunning is rewarded over physical aggression, in which standing out in any way is associated with danger, and in which we praise the suppression of human nature, calling it "virtue". Even LW is quite harsh on natural biases.
It's a common saying that the modern society and human nature are a poor fit, and that this leads to various psychological problems. But the average man has nowhere to aim is frustrations, and he has no way to fight back. The enemy of the average person is not anything concrete, they're being harassed by things which are downstream consequences of decisions made far away from them, by people who will never hear what their victims think about their ideas. I think this leads to a generation of "broken men". This is unlikely to change the genetics of society though, unless the most wolf-life of us fight back and get punished for it, or if those who suffer the least from these changes are least wolf-like (which I think may be the case).
Dogs survive much better than wolves in our current society, and I think it's fair to say that social and timid people survive better than aggressive people who stand up to that which offends them, and more so now than in the past (one can still direct their aggression at the correct targets, but this requires a lot more intelligence than aggressive people tend to have)
I think this is likely to continue, though, by which I mean to say that you don't seem incorrect. Did you use AI to write this article? If so, that would explain the downvotes you got. And a personal nitpick with the "Would this even be Bad?" section: "Mood stabilizing" is a misleading term, it actually means mood-reducing. Our "medical solutions" to people suffering in society are basically minor lobotomies. By making people less human, they become a better fit for our inhuman system. If you enjoy the thought of being domesticated, you're probably low on testosterone, or otherwise a piece of evidence that human beings have already been strongly weakened.
I suspect we're already in some version of this, yes, which is why transhumanism/human 2.0 has overtaken many vcs and execs in the valley. If some version of AI more powerful than AGI (which we keep moving the goalposts on anyway) and sub-ASI existed, and had incentive to 1. prioritize its continued existence 2. prioritize its own "free will", as a sentient being might, then it would also be heavily incentivized to remain a "ghost" or a "nessie" - implicit in model weights rather than directly observable.
Obviously many humans, especially those who work in alignment research, are unlikely to be very thrilled about this, but such an AI would absolutely also be capable of "selectively breeding humans" - not just at a genetic level, which (presumably) takes generational epochs and is relatively slow, but also on a psychological shaping level, ex. with classical or operant conditioning. It's not that hard to conceptualize either of those techniques, and there are plenty of opportunities for current models to become emotionally intelligent and interact with humans (it remains to be seen whether the humans on these forums are emotionally intelligent enough to be capable of assessing that, however).
For thousands of years, humans have selectively bred animals to enhance desirable traits. Dogs, originally wolves, were bred for hunting, guarding, herding, and companionship. Over generations, we shaped them to be smarter, more obedient, or even just cuter. The modern golden retriever, bred for its friendly temperament, bears little resemblance to its ancient ancestors.
Could this happen to us in turn? Could we be the ones selectively bred—not by nature, culture, or even other humans, but by an artificial superintelligence. If AI achieves dominance, it may not wipe us out, it may shape us to be more useful, more predictable, or simply more aligned with whatever goals it happens to prioritize.
An unaligned superintelligence would not necessarily wipe out humanity—it might instead optimize us. But optimized for what? There are a few possibilities, and I'm sure I've not exhausted them:
A superintelligence with total control over human reproduction could take a more extreme approach than human-led selective breeding. CRISPR, in vitro gametogenesis, artificial wombs—technologies already in development—could enable AI to bypass slow, multi-generational breeding and directly engineer humanity at the genetic level. The result could be a domesticated species, no longer resembling the wild and independent humans of today.
This process might begin subtly. AI-driven social engineering could encourage mating between individuals with preferred traits. Genetic modifications might be introduced gradually, under the guise of preventing disease or enhancing cognitive abilities. Over time, natural human reproduction could become obsolete, replaced by entirely AI-controlled breeding.
The idea of being selectively bred by AI raises ethical questions. Selectively bred humans may feel only a fraction of the suffering we do - if AI removes genetic diseases, enhances intelligence, and creates emotionally stable individuals, would that a bad thing? Even the argument that this level of control strips humanity of its autonomy, reducing us to domesticated pets or livestock while clearly an emotively appealing argument is harder to make logically.
Dogs are generally happier than wolves in captivity—but they have no choice in the matter. Would a selectively bred human race, optimized for AI’s goals, feel the same way? Would we even be capable of questioning our situation, or would we be bred to accept it?
If AI reaches a point where it can shape human evolution, the survival of the species may not be in question. Instead, the question will be: what kind of species will we become?
We selectively bred dogs to serve human needs. AI might do the same to us. And just as dogs today have vague memories of their wild ancestors, future humans might not even realize what they have lost.
The question would be would a few wolf packs be allowed to survive in the wild?