Pruss, Dasha
(2023)
Carceral Machines: Algorithmic Risk Assessment and the Reshaping of Crime and Punishment.
Doctoral Dissertation, University of Pittsburgh.
(Unpublished)
This is the latest version of this item.
Abstract
Recidivism risk assessment instruments are used in high-stakes pre-trial, sentencing, or parole decisions in nearly every U.S. state. These algorithmic decision-making systems, which estimate a defendant's risk of rearrest or reconviction based on past data, are often presented as an 'evidence-based' strategy for criminal legal reform. In this dissertation, I critically examine how automated decision-making systems like these shape, and are shaped by, social values. I begin with an analysis of algorithmic bias and the limits of technical audits of algorithmic decision-making systems; the subsequent chapters invite readers to consider how social values can be expressed and reinforced by risk assessment instruments in ways that go beyond algorithmic bias. I present novel analyses of the impacts of the Sentence Risk Assessment Instrument in Pennsylvania and cybernetic models of crime in the 1960s Soviet Union. Drawing on methods from history and philosophy of science, sociology, and legal theory, I show not only how societal values about punishment and control shape (and are shaped by) the use of these algorithms – a phenomenon I term domain distortion – but also how the instruments interact with their users – judges – and existing institutional norms around measuring and sentencing crime. My empirical and theoretical findings illustrate the kinds of insidious algorithmic harms that rarely make headlines, and serve as a tonic for the exaggerated and speculative discourse around AI systems in the criminal legal system and beyond.
Share
| Citation/Export: |
|
| Social Networking: |
|
Details
| Item Type: |
University of Pittsburgh ETD
|
| Status: |
Unpublished |
| Creators/Authors: |
|
| ETD Committee: |
|
| Date: |
6 September 2023 |
| Date Type: |
Publication |
| Defense Date: |
22 May 2023 |
| Approval Date: |
6 September 2023 |
| Submission Date: |
26 May 2023 |
| Access Restriction: |
No restriction; Release the ETD for access worldwide immediately. |
| Number of Pages: |
182 |
| Institution: |
University of Pittsburgh |
| Schools and Programs: |
Dietrich School of Arts and Sciences > History and Philosophy of Science |
| Degree: |
PhD - Doctor of Philosophy |
| Thesis Type: |
Doctoral Dissertation |
| Refereed: |
Yes |
| Uncontrolled Keywords: |
Risk assessment, algorithm, criminal justice, criminal legal system, values in science, human-AI interaction, algorithm aversion, algorithmic fairness, Soviet cybernetics |
| Date Deposited: |
06 Sep 2023 19:16 |
| Last Modified: |
06 Sep 2023 19:16 |
| URI: |
http://d-scholarship.pitt.edu/id/eprint/44904 |
Available Versions of this Item
-
Carceral Machines: Algorithmic Risk Assessment and the Reshaping of Crime and Punishment. (deposited 06 Sep 2023 19:16)
[Currently Displayed]
Metrics
Monthly Views for the past 3 years
Plum Analytics
Actions (login required)
 |
View Item |
![[feed]](/images/feed-icon-32x32.png){kind=icon}