CEU Electronic Theses and Dissertations, 2025
| Author | Sivri, Aysegul |
|---|---|
| Title | AI-Assisted Judging: Legal, Ethical, And Cognitive Implications Of Ai Nudges In The Judiciary |
| Summary | Role of AI has been increasing recently in our judicial systems and its role in shaping legal reasoning requires immediate examination. AI-powered decision-support systems that cannot be considered as neural tools can influence judicial cognition through mechanisms such as anchoring bias, automation bias, and framing effects. This thesis investigates the extent to which these AI-generated “nudges” impact judges’ reasoning and the resulting legal, ethical, and procedural implications. Through a multidisciplinary lens that combines legal analysis, cognitive psychology, and comparative study, the research addresses how AI tools structure legal information, interact with human cognitive heuristics, and potentially alter the dynamics of judicial discretion. Special attention is paid to AI’s deployment in sentencing and parole decisions, particularly risk-assessment tools like COMPAS, and to the broader normative challenges posed by opaque, non-explainable AI systems. The thesis further examines how different legal systems, including those of the U.S., European Union, and China, regulate AI in the judiciary, and proposes best practices to ensure that AI serves to augment, rather than erode, the human-centered foundations of justice. In the end, this study argues that AI systems should be designed and regulated not merely for efficiency, but with careful attention to transparency, explainability, and judicial autonomy. The findings aim to contribute to the development of policy frameworks that preserve the integrity and legitimacy of judicial decision-making in the age of algorithmic governance. |
| Supervisor | Tommaso Soave |
| Department | Legal Studies LLM |
| Full text | https://www.etd.ceu.edu/2025/sivri_aysegul.pdf |
Visit the CEU Library.
© 2007-2021, Central European University