Mike...I can help since I am working on exactly the types of systems that they're like looking at implementing in various government's administrational duties in government and business.
First, AI integration does not mean hand everything over to AI and let it make decisions and control everything the universe. It simply means utilising AI to track and read data, and provide information, to make low level decisions such as scheduling, or to innitiate various types of alerts relating to a particular matter. The functions are basically endless, and they don't require complete control.
I'll give you an example. Let's use technician scheduling and fleet management as an example...
Say you want to use AI to for fleet and scheduling management for a team that does repairs to low income government housing. Instead of some guy sitting there looking at logbooks and making ticks and crosses in boxes, here is what AI can do...
Let's assume the vehicles have onboard computers like most have these days. That data from that computer can be plugged into AI and it will alert the fleet management team not only about problems that have occurred, like a breakdown, but also about maintenance that might be required. It will be able to automatically schedule vehicle service visits and then review service reports to ensure that the service carried out was in line with the manufacturer instructions.
The scheduling for the technicians can be handled by AI in that AI can read that data on the expected time the work will take and slot new "jobs" into the schedule. The AI can be plugged into traffic data from Google Maps, so can make live adjustments to scheduling. It can quickly coordinate replacement vehicles in the event of a breakdown. It is able to seamlessly route vehicles from destination to destination in the most efficient way possible...again, instead of a person sitting there and phoning drivers to find out how long they will be and if they can swing past another job on the way back to the office. And it can alert clients in timely fashion that somebody is coming to their site...and re-alert the office if there is no response from the client, and then it can assist the client in making a new booking...ALL without wasting anyone's time.
If the AI has information on each driver/technician and their skills and experience, it can schedule the correct person to visit particular job. And it can determine whether the guy will need help, and if so it can schedule team members to assist and prompt them in the event that they might need to bring particular tools or materials to the technician they will be assisting.
All of these things still need to be looked at by management, and signed off on. But it increases efficiency massively because the hard work and thinking is done by AI.
All the AI is doing is to make the best use of the data that is available in order promote efficiency...which saves costs.
If I have you 50 drivers/technicians with 100 jobs to be completed in a day, along with the addresses, description of the work to be carried out and a detailed profile of each technicians abilities, AI would carry out tailored scheduling for that fleet of technicians in about a minute. And it would then live-monitors that fleet, communicate with technicians and continue to adjust and amend the scheduling and routing through the day, alerting each technician in real time if when adjustments are made that are relative to them.
...and then it's up to the manager to look over it and see if there is any reason he can think of why the schedule might need adjusting.
That's just one small function that Ai could easily carry out.
So, no, Mike, you don't have to worry that government will hand over all the power to Ai.
All that is happening is that some responsible governments and institutions are realising that focused AIs, created for specific tasks, could massively improve performance and save the public money.