Kamil Mamak
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Kamil Mamak has a total of 29 co-authors. Here are the top co-authors by number of shared publications:
Daniel Kwiatkowski
Jagiellonian University, Poland
3 shared publications- Agnieszka Barczak-Oplustil
Katedra Prawa Karnego, Wydział Prawa i Administracji, Uniwersytet Jagielloński w Krakowie
2 shared publications - Hayate Shimizu
Hokkaido University, Sapporo, Japan
2 shared publications Jadwiga Glanc
Independent scholar
2 shared publicationsJoanna Dudek
Jagiellonian University, Poland
2 shared publicationsMaciej Koniewski
Jagiellonian University, Poland
2 shared publications
Latest Publications
- 11 November 2026
- Published by Taylor & Francis
- p. 47-62
Theory of Punishment and New Scenarios in the Age of AI
This chapter explores the theory of punishment, considering the new challenges posed by artificial intelligence in criminal law. We examine two distinct but interconnected scenarios: AI as the perpetrator of crime and the human offender who commits a crime with the aid of AI. For the first case, we investigate whether retributive punishment can be meaningfully applied to non-human entities, assessing both its theoretical foundations and the implications of extending legal personhood to AI systems. For the second case, we analyse the rationale for punishment when human conduct is mediated, facilitated, or amplified by AI systems, with particular attention to rehabilitative goals. By comparing retributive and rehabilitative perspectives, the chapter proposes a differentiated approach to punishment that accounts for the ontological status of the offender, the social meaning of condemnation, and the preventive needs of the criminal justice system. Ultimately, the discussion explores how AI compels a rethinking of traditional punitive theories, offering new scenarios in which criminal law must balance responsibility, fairness, and technological innovation.
- 11 November 2026
- Published by Taylor & Francis
- p. 47-62
Theory of Punishment and New Scenarios in the Age of AI
This chapter explores the theory of punishment, considering the new challenges posed by artificial intelligence in criminal law. We examine two distinct but interconnected scenarios: AI as the perpetrator of crime and the human offender who commits a crime with the aid of AI. For the first case, we investigate whether retributive punishment can be meaningfully applied to non-human entities, assessing both its theoretical foundations and the implications of extending legal personhood to AI systems. For the second case, we analyse the rationale for punishment when human conduct is mediated, facilitated, or amplified by AI systems, with particular attention to rehabilitative goals. By comparing retributive and rehabilitative perspectives, the chapter proposes a differentiated approach to punishment that accounts for the ontological status of the offender, the social meaning of condemnation, and the preventive needs of the criminal justice system. Ultimately, the discussion explores how AI compels a rethinking of traditional punitive theories, offering new scenarios in which criminal law must balance responsibility, fairness, and technological innovation.
- Open Access
- 20 August 2026
- Published by Taylor & Francis in Inquiry
- Vol. 69 (10) , 3690-3713
AI agency and criminal responsibility: a category mistake
1. The deployment of AI systems has had a significant impact on responsibility which is central ethical issue related to AI (see. e.g. Müller 2020; Gordon and Nyholm 2021). AI could lead to a so-ca...
- 20 August 2026
- Published by Taylor & Francis in Inquiry
- Vol. 69 (10) , 3690-3713
Open AccessAI agency and criminal responsibility: a category mistake
1. The deployment of AI systems has had a significant impact on responsibility which is central ethical issue related to AI (see. e.g. Müller 2020; Gordon and Nyholm 2021). AI could lead to a so-ca...
- 20 June 2026
- Published by Springer Nature in Asian Journal of Philosophy
- Vol. 5 (2) , 57
Who belongs in the moral circle? AI systems and the limits of inclusion
- 20 June 2026
- Published by Springer Nature in Asian Journal of Philosophy
- Vol. 5 (2) , 57
Who belongs in the moral circle? AI systems and the limits of inclusion
- Open Access
- 3 June 2026
- Published by Elsevier in SSRN Electronic Journal
Generating and Verifying Expert-Level Legal Reasoning: A Proof-of-Concept Study in Polish Criminal Law
LLMs are increasingly likely to be used in legal work, including criminal-law settings where unsupported reasoning may have serious consequences. Existing legal benchmarks and exam-style evaluations suggest that LLMs can perform well on some tasks, but do not adequately test expert-level, open-ended, jurisdiction-specific legal analysis. They also tell us little about whether such systems can reliably assess the quality of their own outputs in agentic workflows. To address these gaps, we designed an expert-authored Polish criminal-law vignette and a 54-item rubric and evaluated four LLMs together with a multi-agent pipeline. This paper presents the results of two experiments: case-solving performance and verifier performance. We find that chain-of-thought prompting substantially improves answer generation, with GPT CoT and Claude CoT achieving the strongest results, while the best multi-agent run is competitive but does not exceed the strongest CoT systems. By contrast, verifier performance lags behind generation: apart from GPT CoT and Claude CoT, most systems assess their own answers too leniently relative to expert judgment. We therefore argue that solving a legal case and reliably evaluating such a solution are distinct capabilities. This distinction matters for agentic legal systems, where a weak verifier may reinforce rather than correct erroneous output. Although preliminary and limited to a single vignette, this study shows the value of expert-authored, rubric-based evaluation for legal AI and highlights verifier reliability as an important bottleneck for agentic legal systems.
- 3 June 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessGenerating and Verifying Expert-Level Legal Reasoning: A Proof-of-Concept Study in Polish Criminal Law
LLMs are increasingly likely to be used in legal work, including criminal-law settings where unsupported reasoning may have serious consequences. Existing legal benchmarks and exam-style evaluations suggest that LLMs can perform well on some tasks, but do not adequately test expert-level, open-ended, jurisdiction-specific legal analysis. They also tell us little about whether such systems can reliably assess the quality of their own outputs in agentic workflows. To address these gaps, we designed an expert-authored Polish criminal-law vignette and a 54-item rubric and evaluated four LLMs together with a multi-agent pipeline. This paper presents the results of two experiments: case-solving performance and verifier performance. We find that chain-of-thought prompting substantially improves answer generation, with GPT CoT and Claude CoT achieving the strongest results, while the best multi-agent run is competitive but does not exceed the strongest CoT systems. By contrast, verifier performance lags behind generation: apart from GPT CoT and Claude CoT, most systems assess their own answers too leniently relative to expert judgment. We therefore argue that solving a legal case and reliably evaluating such a solution are distinct capabilities. This distinction matters for agentic legal systems, where a weak verifier may reinforce rather than correct erroneous output. Although preliminary and limited to a single vignette, this study shows the value of expert-authored, rubric-based evaluation for legal AI and highlights verifier reliability as an important bottleneck for agentic legal systems.
- Open Access
- 11 April 2026
- Published by Springer Nature in Science and Engineering Ethics
- Vol. 32 (3)
Ethics of the Design of Animal-like Robots
In the ethical literature on robot design, considerable attention has been devoted to the risks that human-like robots pose to humans. In particular, the resemblance of robots to humans has been associated with concerns about social disruption, deception, and potential harm, leading some authors to argue for restrictions or bans in certain contexts. However, while the ethical significance of human-like form has been widely recognized for humans, significantly less attention has been paid to the implications of animal-like robots from the perspective of nonhuman animals and human–animal relations. This paper addresses that gap by examining how the design and deployment of animal-like robots may affect animals both directly and indirectly. We argue that the use of animal form in robotics is not ethically neutral, as it can shape human perceptions of animals, influence patterns of interaction, and potentially contribute to harms such as mislearning, objectification, and alienation. By shifting the focus from human-centered concerns to the impact on animals, we identify a set of ethically relevant risks associated with animal-like robots and analyze their broader implications for human–animal relations. On this basis, we propose preliminary guidelines aimed at supporting the responsible design and use of such technologies.
- 11 April 2026
- Published by Springer Nature in Science and Engineering Ethics
- Vol. 32 (3)
Open AccessEthics of the Design of Animal-like Robots
In the ethical literature on robot design, considerable attention has been devoted to the risks that human-like robots pose to humans. In particular, the resemblance of robots to humans has been associated with concerns about social disruption, deception, and potential harm, leading some authors to argue for restrictions or bans in certain contexts. However, while the ethical significance of human-like form has been widely recognized for humans, significantly less attention has been paid to the implications of animal-like robots from the perspective of nonhuman animals and human–animal relations. This paper addresses that gap by examining how the design and deployment of animal-like robots may affect animals both directly and indirectly. We argue that the use of animal form in robotics is not ethically neutral, as it can shape human perceptions of animals, influence patterns of interaction, and potentially contribute to harms such as mislearning, objectification, and alienation. By shifting the focus from human-centered concerns to the impact on animals, we identify a set of ethically relevant risks associated with animal-like robots and analyze their broader implications for human–animal relations. On this basis, we propose preliminary guidelines aimed at supporting the responsible design and use of such technologies.
- Open Access
- 3 March 2026
- Published by Cambridge University Press (CUP) in Episteme
- p. 1-15
Disinformation and Superspreaders: With Epistemic Power Should Come Criminal Responsibility
Despite years of efforts to combat disinformation, we remain far from a satisfactory set of solutions. The rise of generative AI, which enables the creation of highly credible fake content at scale, suggests that the problem is likely to grow even more severe. Lessons from the recent pandemic also call for a reconsideration of how disinformation should be addressed. This paper proposes a new approach that focuses not only on regulating everyone who spreads false information, but also on those who hold epistemic power – individuals with the capacity to shape what others know or believe. Such a strategy has the potential to move the debate forward, as it avoids the most common objection to disinformation regulation: the fear of widespread censorship. The paper argues that an individual’s epistemic position can justifiably differentiate their legal duties, and that those who possess epistemic power should bear corresponding legal – specifically, criminal – responsibility for the abuse of that power in spreading disinformation.
- 3 March 2026
- Published by Cambridge University Press (CUP) in Episteme
- p. 1-15
Open AccessDisinformation and Superspreaders: With Epistemic Power Should Come Criminal Responsibility
Despite years of efforts to combat disinformation, we remain far from a satisfactory set of solutions. The rise of generative AI, which enables the creation of highly credible fake content at scale, suggests that the problem is likely to grow even more severe. Lessons from the recent pandemic also call for a reconsideration of how disinformation should be addressed. This paper proposes a new approach that focuses not only on regulating everyone who spreads false information, but also on those who hold epistemic power – individuals with the capacity to shape what others know or believe. Such a strategy has the potential to move the debate forward, as it avoids the most common objection to disinformation regulation: the fear of widespread censorship. The paper argues that an individual’s epistemic position can justifiably differentiate their legal duties, and that those who possess epistemic power should bear corresponding legal – specifically, criminal – responsibility for the abuse of that power in spreading disinformation.
- 14 February 2026
- Published by Springer Nature in Philosophical Studies
- p. 1-24
In defense of artificial suffering
- 14 February 2026
- Published by Springer Nature in Philosophical Studies
- p. 1-24
In defense of artificial suffering
- 14 January 2026
- Published by Springer Nature in Topoi
- p. 1-11
AI Animal Communication: A Potential Source of Moral Revolution
- 14 January 2026
- Published by Springer Nature in Topoi
- p. 1-11
AI Animal Communication: A Potential Source of Moral Revolution
- Open Access
- 13 December 2025
- Published by Elsevier in Journal of Criminal Justice
- Vol. 102
Are sentencing disparities unavoidable? Lessons from centralized judicial training
- 13 December 2025
- Published by Elsevier in Journal of Criminal Justice
- Vol. 102
Open AccessAre sentencing disparities unavoidable? Lessons from centralized judicial training
- Open Access
- 22 September 2025
- Published by Springer Nature in International Journal of Social Robotics
- Vol. 17 (11) , 2743-2753
Robot Kingdom: On the Difficulty of Finding the Definition of Robots
- 22 September 2025
- Published by Springer Nature in International Journal of Social Robotics
- Vol. 17 (11) , 2743-2753
Open AccessRobot Kingdom: On the Difficulty of Finding the Definition of Robots