Is it the right move
In his article “Artificial Intelligence in Policing”, Noah Feldman explores an issue few may have heard about. It seems that the New Orlean Police Department has been using Artificial Intelligence (AI) to help it fight crime.
To some, given the positive contributions made by AI, this seems to be a good idea. Unfortunately, there are a lot of dangers and issues that need to be overcome before I can effectively become a policing agent.
Prediction is not evidence
Science is built on the idea that predictions drive scientific exploration, but predictions are not infallible nor universal. The legal system uses evidence, proof, justice and even mercy to achieve its goals. These latter characteristics are supposed to be a foundation for police work as well.
The vulnerability of prediction is that it may think they have picked an area where crime will occur. Yet human nature being what it is may disappoint those predictions. Using AI to predict where crime will occur would also violate the long-held truth that all people are innocent until proven guilty.
Prediction is not infallible
While artificial intelligence could make positive contributions to police work, the fact that predictions are not infallible would cause major resource problems. Mr. Feldman rightly pointed out that these predictions would divert police attention, funds and resources to investigations on innocent people of New Orlean.
Supporters of this contribution may say that the AI police agents would have strong algorithms guiding them. But sadly for them, those algorithms operate under probability, not evidence. The evidence is needed to bring a criminal to trial and convict them.
Probability is not evidence, but an assumption and assumptions do not fly in a court of law.
Justice, mercy and AI
Justice cannot be meted out by a machine. AI is all reality s a machine which does not understand the basics of what is or isn’t just. AI also lacks the ability to apply mercy when mercy is needed. It works on a system of parameters and not human judgment.
The supporters of this New Orlean police use of AI may argue that the ideas of justice can be programmed into AI. While that is a possibility, it is like trying to fit a ‘square peg in a round hole.’ It simply will miss what policing and law enforcement are all about.
In its present form, it is easy to see how AI would be accepted by a small group of people. These are people who disagree with the current definitions of justice, mercy, and right and wrong. And want to implement their own ideas, which so far has been seen in the advent of zero tolerance.
This form of justice does not work as it does not take into consideration the many reasons for why a crime was committed. It also excludes mercy and justice from being part of the process.
AI can make positive contributions to policing, but there are some areas of life that need to be left to human judgment alone. That is if we want a good society to live in.