About The Workshop
SW 104– Ethics, AI, and Criminal Justice
Convenors: Frej Klem Thomsen; Sebastian Jon Holmen
Contact: fkt@dketik.dk
Artificial intelligence is rapidly transforming criminal justice. Predictive analytics, facial recognition, deepfakes, automated decision-support tools, and generative AI are increasingly used—or proposed for use—by police, prosecutors, courts, victims, defendants, and public authorities. These developments raise urgent questions for legal philosophy and social philosophy: How should criminal justice institutions evaluate AI-generated evidence, testimony, or expressive material? What forms of transparency, accountability, and contestability are required when AI systems shape coercive decisions? How should legal systems weigh efficiency and investigative power against equality, privacy, dignity, procedural fairness, and the rule of law?
This workshop addresses these questions through the theme “Ethics, AI, and Criminal Justice.” It is closely aligned with the IVR 2026 congress theme, “Law in the Face of the Changing Problems of the World,” because AI presents a paradigmatic changing problem for contemporary law: it alters not only legal practice, but also the moral and conceptual assumptions on which legal institutions depend. This workshop narrows that wider inquiry to the criminal justice context, where state coercion, punishment, surveillance, and vulnerability make the ethical stakes especially acute.
The workshop as submitted has two confirmed contributions. “The Ethics of AI-generated Victim Impact Statements” examines the 2025 Arizona case in which a deepfake of murder victim Chris Pelkey was submitted as a victim impact statement at sentencing. The paper asks whether AI-generated victim statements are objectionable because of their artificiality, their relation to intention and authorship, or their potential influence on sentencing. It argues that some familiar objections to deepfake victim impact statements are weaker or narrower than often assumed, and that in some cases such statements may even better serve certain functions of victim impact statements.
The second confirmed contribution, “Three Challenges to Police Use of Facial Recognition,” examines facial recognition technology as an AI system used to identify or compare persons through faceprints. The paper focuses on three prominent ethical challenges: undesirable deterrence effects, algorithmic bias, and reductions in citizen privacy, including slippery slope concerns. It argues that public and policy debates often underestimate the complexity of these challenges and therefore mischaracterize their severity or nature.
Together, these papers frame a broader workshop inquiry into the ethical governance of AI in the context of the criminal justice system. The workshop invites additional papers on topics such as AI-generated evidence, deepfakes in criminal proceedings, algorithmic risk assessment, predictive policing, biometric surveillance, automated sentencing tools, AI and victims’ rights, AI and defendants’ rights, explainability and due process, bias and discrimination, human dignity, accountability for AI-assisted criminal justice decisions, and comparative regulatory responses.
As required by conference guidelines, the workshop will be structured to promote discussion. We anticipate one or at most two sessions, depending on the number of accepted papers. Each paper will be presented for approximately 15–20 minutes, followed by at least 20–30 minutes of collective discussion per session. Draft papers, extended abstracts, or presentation outlines will be circulated in advance to enable prepared engagement.
The workshop will be inclusive, interdisciplinary, and international in scope, welcoming contributions from legal philosophy, criminal law theory, ethics, human rights, political philosophy, technology law, criminology, and socio-legal studies. It aims to clarify how criminal justice systems can respond to AI without abandoning core legal values: fairness, accountability, equality, privacy, dignity, public reason, and responsible punishment.
Confirmed speakers:
- Sebastian Holmen, Roskilde University, “The Ethics of AI-generated Victim Impact Statements”
- Frej Klem Thomsen, Danish National Centre for Ethics, “Three Challenges to Police Use of Facial Recognition”
Open call: The convenor(s) welcome additional papers.
Accessibility and format-related needs: No special format-related needs are currently anticipated. The convenor(s) will ask presenters to use accessible slides, speak at a measured pace, describe key visual content, and make materials available in advance where reasonably possible, consistent with the IVR 2026 accessibility guidance.
