I believe AI/ML can reduce government costs and boost efficiency by automating staff duties. But if I leave actual governing to a machine, I am left with an ent

I believe AI/ML can reduce government costs and boost efficiency by automating staff duties. But if I leave actual governing to a machine, I am left with an entirely new, but not necessarily smaller attack surface for corruption. Traditionally: monied interest buys influence, maybe in the form of a steak dinner for the mayor. AI Gov’t Future: monied interest buys asymmetric understanding of GovBot5000, perhaps by hiring an AI researcher or data firm to figure out how to exploit flaws in the algorithm. If we tackles issues of inefficiency, opacity, and corruption in government first, then I think we can tackle judgment and representation, which are less “on fire” things imo.