Law and Artificial Intelligence at a State Level
Algorithms in our legal system are becoming more prominent. Law and artificial intelligence are now forever related. This also creates a lot of questions. Some examples are changes in state law in New Jersey, Arizona, and others.
There has been an effort to install algorithms in sentencing guidelines. And there’s some pros and cons there. The pros are you’re not going to have abhorrent rulings from judges that are outside the bounds of the law. You should have a fair sentence. We should impose them based on facts and law and mitigating factors, for example.
So there should be more, in theory at least, consistency in sentencing. Having said that, the human element of sentencing a criminal defendant is now gone. This is because you don’t have a judge who can check the remorse of the offender. There isn’t the check on the impact upon the victims. There’s no weighing the cost-benefit analysis to incarcerating someone or punishing someone.
Human Perspective Matters
That human perspective is important. Having said that, using algorithms, or AI, for lower level offenses makes good sense. It relieves our overwhelmed court system, and it allows speedy disposition of justice. AI in the criminal justice system could better consider certain factors. The algorithm can make appropriate consideration of age and prior criminal history. So the more that the system can filter these factors, the more consistent sentences there will be. It also helps the criminal defendant have a fairer consideration. The defendant has the benefit of the whole picture for the computer’s consideration. And bear in mind, there is a fail-safe in place, no matter what happens at the judge level or the computer level. There’s a Court of Appeal in every jurisdiction and at the federal level. So there is a certain degree of human oversight over the process.
Law and Artificial Intelligence: Considering Many Factors
At the fundamental level, considering as many factors possible is a benefit. It would allow for just, fair, and expedient sentencing. There are some limitations in implementing those rules. In particular in a way where the automated system is not rolled out too fast and we don’t have issues and errors. [First, using the system for lower level offenses or cases with a lower level dollar value in the civil system makes good sense.] This is because the consequences of an erroneous ruling won’t be grave or profound. There can always be a different way to handle that. There can always be an appeal, for example, of those issues.
So I would start at a low level. In criminal offense would make good sense. This is whether it’s a misdemeanor level or a limited civil case. It’ll allow testing of the system. Afterward, the hope is that we can start applying it to bigger cases. The ultimate goal of the AI is to make things fair, efficient, and consistent, in the legal context. And will give our courts significant relief.
