One way I think we could curb AI wealth inequality is if we regulate that all AI agents or robots (who produce useful work) must be associated with a real human

One way I think we could curb AI wealth inequality is if we regulate that all AI agents or robots (who produce useful work) must be associated with a real human and their fiscal number. If a factory needs 200 droids then 10% of droid generated profit goes to the real human associated with that droid. If a lawyer office needs 200 processes running AI agents searching through common law cases then 200 ppl would have to receive a fee for AI agents' use. You can extrapolate to anything you want but basically the idea that very few humans actually need to work but that all 'work' must be regulated and associated to a real human with a real fiscal entity, not simply a thread or a company asset in the case of a physical robot. In this scenario the question of a single uber human, master of all trades, no longer makes sense.