Comment by dannyobrien

As I implied in an earlier comment, your conviction (if you're wrong on the inevitability of the direction), may be one of the things that leads it into that direction.

Here's my historical take: in the 1960s and 1970s, computation in general was viewed as a sinister, authoritarian thing. Many people assumed it was going to be that way, and a small minority recognised that it also had the potential to empower and grant autonomy to a wider class of people. These were the advocates of the personal computer revolution -- the idea of "computer lib", whereby the tools of control would be inverted and provided to the people at large.

You can argue about whether that strategy was a success or not, but the group tht was largely irrelevant to that fight were the people who decided not to get involved, or to try (although not very hard) to impede the development of computation in general.

To bend the trajectory of AI in general involves understanding and redeploying it, rather than rejecting it. It also involves engaging. If it's anything like the last few times, the group that is simultaneously exploring and attempting to provide agency and autonomy for the maximum number of people will be smaller than both those using new tech to exploit people or maintain an unequal status quo, and the people who have good intentions, but throw their hands up at the possibility of using their skills to seize control of the means that provide for a better future.