Perfection will always elude us.
But the more I think of it, the more I like the idea of an open-source AI being used to scour existing legal documents, plus evidence and testimony, and then rendering a preliminary verdict of “X”, with a full report as to the logic of this recommendation.
Then a human judge (who we hope would be unbiased (!)) could review the verdict and either accept it, or state an argument as to why, based on the evidence, the AI has got it wrong, and state the actual verdict of “Y”, with a full report as to why the human judge feels correct in this judgment.
As part of the appeals process, a higher level court could then review the evidence, review the AI report, and review the official judge’s report. An appeals court could then re-assess.
But I think AIs will soon play an important role in these things. Humans will be a fail-safe, but logic processing would be beneficial.
Could start with Professional sports games and officiating calls. Fans claim they want more perfect calls but I think they really want more favorable calls for their side and overlooking of their teams’ transgressions. I have read credible opinions that there are actionable flaggable penalty offenses on every play in the NFL and NBA. I think the consistency of what percentage are “called” by the refs and what are ignored and the parity of this between teams in each match is the issue. Oh, and cheating is going on; lots of it I think.
An interesting concept and one I hadn’t really considered before; primarily because AI can be biased with initial programming. However, if AI can be what some say it can be, then by it’s nature, it should be able to recognize its own bias and prevent it. But, that is a long, long way off.