Neuroscience, Artificial Intelligence, and the Case Against Solitary Confinement
81 Pages Posted: 3 Jul 2019
Date Written: July 2, 2019
Prolonged solitary confinement remains in widespread use in the United States despite many legal challenges. A difficulty when making the legal case against solitary confinement is proffering sufficiently systematic and precise evidence of the detrimental effects of the practice on inmates’ mental health. Given this need for further evidence, this Article explores how neuroscience and artificial intelligence (AI) might provide new evidence of the effects of solitary confinement on the human brain.
This Article argues that both neuroscience and AI are promising in their potential ability to present courts with new types of evidence on the effects of solitary confinement on inmates’ brain circuitry. But at present, neither field has collected the type of evidence that is likely to tip the scales against solitary confinement and end the practice. This Article concludes that ending the entrenched practice of solitary confinement will likely require both traditional and novel forms of evidence. In exploring the potential effects of neuroscientific evidence on support for solitary confinement, the Article reports results from an original online experiment with a group of 250 ideologically conservative participants. The analysis finds that the introduction of brain injury reduced conservatives’ support for solitary confinement but not to the extent that is likely to make a policy impact. The Article argues that future, more individualized brain evidence may be of greater use, but at present neuroscience is limited in its ability to systematically measure the brain changes that inmates experience in solitary confinement.
This Article then turns to AI and argues that it could be developed to provide litigators and inmates with the ability to more effectively document the detrimental effects of solitary confinement. Looking to the future, the Article lays out a vision for an AI system called “Helios,” named after the Homeric sun god believed to see and hear everything. The Article envisions Helios as a self-learning AI system with a mission to help inmates and their attorneys gather more systematic evidence of the effects of solitary confinement on inmate health. Helios is also a platform on which additional inmate services might one day be provided. The Article describes how Helios must be carefully designed, with particular attention given to privacy concerns.
This Article is organized in seven parts. Part I describes the historical and contemporary use of solitary confinement in the United States, highlighting the known effects of solitary confinement on inmates. Part II summarizes recent constitutional challenges to the practice of solitary confinement. Part III explores the potential for integrating neuroscientific evidence into these legal challenges to solitary confinement. Part IV discusses a new online experiment to explore whether neuroscience might change public opinion on solitary confinement. In Part V, the Article transitions to a consideration of AI. The Article proposes a self-learning system, Helios, and describes how the system would operate. Part VI turns to a series of challenging ethical and legal questions about the design and implementation of Helios. Part VII briefly concludes.
Keywords: law and neuroscience, brain, solitary confinement, artificial intelligence, criminal justice, mental health, inmates, empirical
Suggested Citation: Suggested Citation