Finding a balance: machine learning in bail and bond

Abstract

This paper seeks to assess whether Machine Learning can address the existing gap in the determination of bail and bond by focusing on the causation and correlation of the cognitive bias of judges and the wide-ranging discretionary power they wield. The paper shows that the heavy reliance on pretrial detention occasioned by cognitive biases breaches the rights of accused individuals and heavily impacts their right to a fair trial. This results in subversion of justice, weakening the criminal justice system and, by extension, the rule of law. The paper then argues that machine learning devoid of the corporeal limits of human cognition, can streamline the bail process ensuring that the rights of accused persons are preserved. The paper relies on the principle of fairness to work out the minimums that the Kenyan Law should provide for the integration of machine learning algorithms. The paper contends that players in both the justice and technology sectors need to ensure that the use of machine learning algorithmic systems is well-regulated and in correspondence with the rule of law and the principles of fairness.