Algorithmic Justice: The Legal Implications of AI in Criminal Sentencing and Risk Assessment

Artificial intelligence is becoming deeply embedded in criminal justice systems, particularly for sentencing decisions and risk assessment. Tools like COMPAS and HART promise to make the justice system more consistent, efficient and objective, yet they bring serious legal and ethical problems. The algorithms typically draw on historical data that contains existing biases, producing discriminatory results that disproportionately affect ethnic minorities and reinforcing the very inequalities they claim to address. Many of these systems are proprietary black boxes that courts and defendants cannot examine or challenge, undermining the transparency and accountability required for fair trials. This article analyses the legal consequences of using algorithmic risk assessment tools in criminal justice, arguing that adopting them without careful scrutiny threatens core legal protections: the presumption of innocence, the right to a fair trial and the role of judicial discretion. Without serious reforms to address algorithmic bias, ensure transparency and protect due process, flawed computer predictions risk replacing human judgment in determining people’s fates.