But - I'd like to encourage you for asking. That's a really good question. If more tech-minded folk asked questions like this, we'd be doing a whole lot better as a species.
My personal hot-take is that understanding how people get manipulated and coerced into acting against their own interests is essential to our survival. Think big tobacco tactics, big oil climate denial, manufacture of consent, kayfabe politics.
I don't know how - or even if - AI can help us create systems that offer protection against such manipulation. My hunch is that it's worth thinking about.
Some examples: Maybe AI can analyze vast amounts of data to manage pollution - soil, air, water, etc - in such a way that manipulation of the results is impossible. However, the institutions which manage this stuff seem quite thoroughly broken (Superfund sites, for example, or Deepwater Horizon, or pesticide buildup).
Maybe AI can analyze currency transactions to identify bribes, sinks in the money system where value flows out. However, we've seen many examples of this stuff coming out into the open without major consequence (Paradise, Panama, bank bailouts, CDS, etc etc).
Maybe AI could help identify those most susceptible to manipulation in order to help & protect them, rather than to sway their vote, turn them into 'useful idiots', and sell them stuff. That would be "bad" for the economy (and fantastic for society).
Maybe AI could help overhaul education - I think this is both essential, and terrifying. I won't ramble on though.
Best of luck man.
Does identifying as an AI expert preclude starting small?
Good luck.
You want to help humanity? Find a way to kill off advertising. It's the cancer that's not just funding the current downsides of technology, but reshaping the technology industry to serve it even more.