Predicting Proportionality: The Case for Algorithmic Sentencing

Expert system gradient descent style training: Development of a defensible artificial intelligence technique

“…Artificial intelligence systems have found use in numerous areas of society. These systems are used to screen loan applicants [1], to make sentencing recommendations for criminal defendants [2][3][4], to scan social media posts for disallowed content [5] and identify poster mental health issues [6], to make medical recommendations [7] and more. Many systems are designed with a capability to learn from the data presented to them as part of a supervised [8], semi-supervised [9] or unsupervised [10] training process.…”

Law, liberty and technology: criminal justice in the context of smart machines

“…In a paper that Vincent Chiao (2019) gave at the Transnational Law Summit that was held at King's College London in April 2018, it was proposed that AI might be used to guide judges in exercising their sentencing discretion, particularly to guide judges as to the proportionality of their dispositions. The thinking was that, where judges have considerable discretion as to both the type of disposition and the scale of the penalty (such as the length of a custodial sentence), it might be helpful to be aware of the ‘norm’ for a case of the kind at issue.…”

Gradient descent training expert system

“…Artificial intelligence and machine learning systems have found use in numerous areas of society. They are used for applicant screening [1], making medical recommendations [2] and even making sentencing recommendations for those convicted of crimes [3][4][5]. However, despite -or perhaps due to -the multitude of areas that these systems are used in, the public is concerned about them [6].…”