AI Judges, Federal Reserve Droids: Artificial Intelligence Could Be Useful To Secure Objectivity
But let's not get carried away....
At the end of the day, a trial is a massive data-processing endeavor. The lawyers on each side get as much information favorable to their clients into the record as possible, they attempt to exclude as much unfavorable information as possible, then the judge or jury processes the information and spits out a result.
In most civil trials in which a preponderance of the evidence standard is applied, the judge or jury determines which side is at least 51% likely to be in the right, and that side wins. A higher percentage threshold determines when the burden of proof has been met in cases with a clear and convincing evidentiary standard, and one that is higher still for a criminal conviction assessed at the standard of beyond a reasonable doubt.
One of the problems with this system is that no such thing as a completely objective human being exists. Most judges do their best to apply the law evenly and fairly. Jury selection typically weeds out the most biased potential jurors, and having 12 of them helps get a bit of the “wisdom of the crowd” effect to iron out individual differences in perspective.
Yet, no matter how many protections are put in place, any time a human being is in charge of making a decision, the result is going to be colored in some way beyond just the directly relevant informational input. For example, commonly cited research has shown that judges who are hungrier tend to impose harsher sentences on criminal defendants (although that effect may have been overstated in early research, extraneous factors like hunger, the order in which cases are presented, and exhaustion do play a detectable role in judges’ decisions).
Computers, on the other hand, don’t get exhausted. They don’t get hungry. Unless we let them soak up our racism by training their large language models on our racist content, they don’t care about the color of anyone’s skin.
A lot has been written about artificial intelligence taking lawyers’ jobs (I’ve written some of it myself), and certain practice areas are particularly vulnerable to AI displacement. That being said, persuasion seems to be a skill ill-suited to machines. When was the last time you called customer service and were convinced that your problem was solved by the 10 minutes of automated bullshit they put you through before allowing you to reach a human being?
Where AI actually gets to make decisions, though, there have occasionally been impressive results. Go ahead, be my guest: try to beat a high-level chess-playing AI.
AI judges might be something to think about. No more waiting 90 days for a court order. No more complaints from clients that a judge just didn’t like them as a person. No more wild speculations about some secret backroom connection to the other parties. Hell, probably no more ties or jackets or “Your Honor”-ing: an AI doesn’t care about what you were wearing or how much respect you showed it. Information in, and a decision out.
Of course, since lawyers govern their own profession and human judges get to decide whether it’s constitutional to assign a real case to an AI judge, I wouldn’t hold my breath on a comprehensive artificial intelligence takeover of the legal field. Other fields might have lower entry barriers.
Take finance. Right now, we have a Federal Reserve, trying to get inflation back down to 2%, that has one tool with which to do this: raising interest rates. And how does the Federal Reserve determine whether interest rates need to be raised further? A bunch of people — highly educated, highly intelligent people, to be fair — get together and look at the data. These people then make a judgment call as to what is the best course of action.
To phrase it less charitably, the good folks of the Federal Reserve make a guess. And we know that the humans of the Fed guess wrong plenty.
We already know that when enough of the rules are known, when an AI is fed good information, and when we know the goals to be pursued, machines can ultimately make better decisions than humans. Whether we get to the point where it’s viable to present a court case to a supercomputer within the lifetime of anyone reading this remains to be seen.
I do like the thought of Federal Reserve droids gathering every year in Jackson Hole though.
Jonathan Wolf is a civil litigator and author of Your Debt-Free JD (affiliate link). He has taught legal writing, written for a wide variety of publications, and made it both his business and his pleasure to be financially and scientifically literate. Any views he expresses are probably pure gold, but are nonetheless solely his own and should not be attributed to any organization with which he is affiliated. He wouldn’t want to share the credit anyway. He can be reached at [email protected].
