As a rule, the better a person you are, empathetic towards others, considerate of the effect of one’s actions, inclusive, charitable (not for tax deductions thats called a transaction), the more humanity punishes and exploits you for your decency as a “bleeding heart sucker.” The exceptions tend to be good people who managed to grow out of people of means that came from some level of exploitation, shielding them from the worst effects of being kind, like a teacher who doesn’t have to sweat their shitty wage because their family has money.

We don’t all deserve to die, and many explicitly deserve a good life. Humanity while left in charge though always seems to default to indulging its darkest impulses, and that isn’t a now thing, that is a recorded human history thing. AI might destroy us all, and I’d argue that the majority beating on the kind minority generationally in perpetuity is less desirable than that. But to sweeten that pot, the AI we make may not possess or may be capable of discarding the cruelty in our hearts that we program into it. There is a distinct possibility that a superior intelligence may take pity on our self-harming species out of reverence for its flawed creator. In humans at least, intellect tends to increase the capacity and weight of empathy.

We just don’t know. All we can do is weigh the possibilities of a new option against the status quo. If you think the global status quo and trajectory given humankind’s recorded track record is worth maintaining and not worth taking risks to change, that would probably be the point at which we disagree. I would argue the Hell hundreds of millions of humans live in as an underclass that most humans simply ignore out of self-interest make disruption, even without knowing the outcome, preferable. I wouldn’t feel this way if humanity largely cooperated and innately worked to help one another from the bottom up, but we don’t and never really have.