We, at Kiiji Universe, have created an app that uses an AI algorithm that can predict, with great accuracy, the recidivism risk of a prison resident before they are released back into society.

​

We understand that many companies and organizations have been trying to develop an AI program that can predict recidivism for individuals already released into society but could not seem to get around the biasness that automatically comes when "race" is introduced as a training feature. This is because AI has not yet learned to discern that, for example, a black person is more likely to be stopped and searched, than a white person, simply because he's black. All it knows is that it happens.

​

Phorsight was created to be used prior to an individual being released back into society. This allows us to alleviate biases by focusing only on an individual's performance while incarcerated. The features used in our program are based on a national study of the most prominent factors affecting recidivism risk.

For example, a prison official may suggest including the number of behavior modification programs completed. However, research indicates that after the 1994 crime bill removed free college courses from prisons, recidivism rates increased from 65% to 72%. Studies show that higher education or trade training significantly reduce recidivism risk to about 16% or less.

​

This does not imply that behavior modification programs are ineffective but highlights that they are less impactful than education and trade training. We aim to use evidence-based practices to equip individuals with the best support for reintegration into society while minimizing the risk of returning to prison.

​

WARNING: This program was not designed for prison officials to determine if a resident should be released from prison, or not, but rather it is to be used as a tool to help key in on what areas a resident need to work on to better prepare them before they are released.

 

​