Justice measures for AI sentencing

Provides justice measures to be used in algorithms used by AI in legal systems.

Some courts already use AI to provide aid in the liklihood of the suspect to appear before court again and commit another crime. In some situations, AI recommendations are used in determining whether the person is to be detained or set free.

This comes with the problem that the algorithms conducted in such AI systems are not trusted. Governments usually buy them instead of developing these algorithms by themselves. Which leads to the outcome that the decisions on individuals made by courts are actually determined by the corporates who make such systems, instead by the courts’ own judges. The factors upon which the AI make a decision are not known by the public and maybe even by the judges who use this system.

This solution aims to develop certain standards for measures to be used by AI systems in the application of justice. Those measures should be categorized, detailed, based upon different prospectives and more importantly should be open. 

The digital transformation and the use of AI are going to dominate the legal field as they will in most, if not all, fields, however this solution aims to give judges control of technology instead of replacing them with it. AI, by this solution, will be a tool in aiding judges to practice their discretion. Different ‘measures of justice’ that AI systems use will be transparent to the judges who can control how to use and apply them.

Judges, as well as others, can also be involved in the development of these measures. This solution will welcome the feedback of judges, legal academics, legal practitioners as well as others (e.g. human rights organizations) in adding measures and the data to be analyzed. This data could be based on precedents or on the individual history.

The following Wired article mention one face of the problem that this solution aims to address.

https://www.wired.com/2017/04/courts-using-ai-sentence-criminals-must-stop-now/

While courts in some states already use algorithms to help decide in sentencing, these systems are flawed as they let corporations made such judicial decisions without even allowing the judges to know the reasoning behind such decisions.

AI systems are not supposed to replace judges but to be used as a tool in their hands to enhance their work.

The proposed solution solves this by making the involvement of judges in making the AI system a crucial part. This is done by making judges contribute to making a set of definitions understandable by a computer system to justice.

While judges play a crucial rule in making the system, this role comes to an end. Judges rule is to make a set of definitions of justice based on data collected on the case in question, similar cases, the suspect and suspect of similar cases. Machine learning experts are needed to take such metrics that define justice and use them in the making of an artificial intelligence system.

1. Making an initial survey to judges on applicable metrics of justice.
2. Building a computer-understandable format with machine learning engineers to fit the metrics of step one.
3. Making continuous sessions with judges for formulating justice measures on the format from step 2.
4. Working with machine learning engineers to build the system with measures from step 3.

By partnering with courts in Egypt and US.

Delending on the populations served by partner courts.

I'm working as a prosecutor for more than eight years. Prosecutors also are part of judiciary in Egypt and become judges at the age of thirty. In addition to that, I will be studying a master of laws in Boston. This gives me

It will allow the solution to solution to make parterships with data science institutions.

Allowing the team to work with partners to conclude the project.