Yes, precisely. That’s why the only approach that might make sense in short to medium term is a system based on individuals bidding. That is, have individual people each estimate the value or cost of laws for themselves. I’m not convinced if the simplest possible system of just simply bidding would work, but I’ll outline it here regardless. Basically, have a platform for the law proposals and bids as well as negative bids. The bids should be in the form of monthly or yearly payments.
The basic premise of the system would be that when positive bidders are offering more money, in total than negative bidders are requesting, then negative bidders will each receive what they asked for, annually and positive bidders will pay for that. In this case, the law becomes eligible for enforcement. If not, no money moves and law doesn’t become eligible for enforcement.
The cost of enforcement also needs to be included in the system somehow. It’d probably work to have positive bidders each define how much of their bid will be applied towards enforcement and how much towards bidding for eligibility.
One potential problem in this system is that if the bids are unbounded, it’s likely that many, if not nearly all, laws would end up uneligible due to a small number of individuals basically vetoing them by applying ridiculous negative bids. I’m not entirely convinced that this would be a problem in practise, but if it is, one potential solution would be to just ignore the highest (for exampel) 5% of negative bids for the specific law. This would discourage people from trying to milk the system as well. However, this would also lead to basically ignoring the costs for laws that only affect under 5% of the population negatively, so better make sure the percentage is set as low as you can without rendering the system useless.
Another thing that could potentially work is to make the identities of the people placing highest negative bids public. In this case the deciding rule should probably be based on the difference from the average or median negative bid. This would allow bidders to at least know who they need to convince. Of course, this is moot if all identities and bids are public anyway, which is worth considering also.
Another potential problem is that some people will see the system merely as a way to get money for nothing. To some degree, this is actually a design goal of the system. It’s a potential way to implement basic income, afterall. However, without a very enlightened populace, it’s likely to become a problem if it becomes popular to place only negative bids. For this reason, I suspect there needs to be an upper limit for how low the sum of all of your bids is allowed to be. In other words, the system doesn’t allow you to get whatever you ask from it, there’s a limit you can’t pass. However, even with this limit, it does allow individuals to “tax” laws they don’t like while funneling the proceeds towards laws they do like.
Richer individuals are likely to see more benefit in paying for laws that help their businesses than taking the basic income for themselves, so they will voluntarily forfeit the basic income in exchange for more voting power.
Anyway, this is a very raw version of an idea of how to implement something like this. The basic premise is to crowdsource the valuation process rathan than trying to build an algorithm to solve it. It’s not quite an AI algorithm, but it is a herding algorithm based on the premise that our pre-existing independent intelligences (whether Artificial or not) are capable of learning how to do the valuation if presented with suitable incentives.