Publications and research outputs.

Hybrid Intention and Augmented Agency: Allocation of Criminal Culpability in Neuroprosthetic Control

Neuroprosthetic systems translate neural or muscular signals into physical action, so harm can follow from conduct that is neither purely manual nor meaningfully autonomous. The article develops hybrid intention for cases in which a human purpose remains identifiable while the system materially shapes how that purpose is carried out, and sets an evidentiary threshold — case-level explainability, prior validation and sufficiently complete logging — that a system-mediated output must pass before it can support attribution. It then locates the control actually available to the human agent across design, configuration, operation and emergency.

From Admissibility to Contestability: Structural Opacity, Encrypted-Platform Evidence, and the Limits of Adversarial Review

Material from encrypted communication platforms now appears in serious criminal cases, while its origin and processing often stay opaque to the defence. The article argues that admissibility alone is too weak a safeguard and that proceedings should secure the practical ability to contest such evidence. It asks what adversarial review requires when the underlying methods are structurally difficult to examine.

Rethinking Human Rights Protection in the Age of AI: From Ex-Post Remedies to Rights-by-Design

Artificial intelligence increasingly mediates decisions that affect rights in areas from social welfare to the judiciary, yet legal protection still tends to react only after harm occurs. The article argues that safeguards should move toward prevention, building rights protection into the design of systems so that problems are caught before they cause harm. It draws on EU instruments, including the AI Act, and cautions that such measures need measurable outcomes and genuine accountability.

Robots as victims? Examining criminal law's boundaries in the digital age

The article asks whether criminal law should treat robots and AI systems as victims of crime or keep its human-centred approach. Reading Australian law alongside developments in the EU, the United States and Japan, it argues that robots should remain classified as property and should not gain legal personhood. It concludes that criminal law can address harms involving robots without extending victim status to autonomous systems.

Referenced in Dmitry V. Zhmurov, Nekriminologiya (2026), in the chapter on robots, artificial intelligence and robot victimology.

A Critical Review of the Institute of Multiple Recidivism in the Modern Criminal Law of the Republic of Serbia: The Controversy of the Current Legal Solution and Possible Solutions

The article examines the reintroduced rule on multiple recidivism in Serbian criminal law, which raises sentences for repeat intentional offenders. It argues that the current threshold leaves courts little room to weigh the circumstances of individual cases and questions whether the rule is justified for most offenders. It suggests that alternative measures may often be more appropriate and notes that the provision is likely to be revised.

Sporazum o troškovima krivičnog postupka – pojam, sporna pitanja i moguće procesne zloupotrebe

The article analyses the agreement on the costs of criminal proceedings as a mandatory component of Serbian plea and testimony agreements, identifying legislative gaps around the injured party's right of appeal and the relief of the defendant from those costs.

Hybrid Intention in Criminal Law: Rethinking Culpability in Posthuman Contexts

This article develops the concept of hybrid intention as a framework for attributing criminal responsibility in human-AI systems. It argues that culpability can be assessed where human initiation, technological mediation and a remaining control window are present, and formulates a three-part attribution test for posthuman contexts of criminal law.

Од пеналног популизма до номотехничке корекције: Члан 55а КЗ као тест пропорционалности и судијске дискреције у Нацрту ЗИД КЗ

On Article 55a of the Serbian Criminal Code as a test of proportionality and judicial discretion in the Draft Act amending the Criminal Code.

Члан 55а Кривичног законика: Генеза, међународна поређења и потреба за изменама

On the genesis of Article 55a of the Serbian Criminal Code, international comparisons and the case for amendment.

AI-Driven Criminal Justice: The Sharp Edge of the Automated State

The contribution examines predictive policing, risk assessment, opaque courtroom tools, and the limits of judge-in-the-loop oversight, arguing that automated evidence should not determine legal outcomes unless it can be meaningfully contested.

Veštačka inteligencija u fudbalu: krivičnopravni izazovi i perspektive

The chapter examines criminal-law responsibility where artificial intelligence is used in football — from VAR and automated refereeing to the detection of match-fixing — and asks how liability should be distributed among referees, federations and software producers.

Pojam i praktične implikacije veštačke inteligencije

The chapter works from a conceptual definition of artificial intelligence toward the mechanisms that let such systems operate more or less autonomously, and closes on the general social benefits and risks that follow.

Neka pitanja u vezi sa zaštitom prava na suđenje u razumnom roku

The paper addresses contested questions in the protection of the right to a trial within a reasonable time, particularly in enforcement proceedings conducted by public bailiffs rather than courts, and the State's liability for the resulting damage.

Privatni zatvori u Sjedinjenim Američkim Državama

The chapter traces the development of private prisons in the United States from the earliest involvement of the private sector in the execution of prison sentences to current models, and considers the effect of the Covid-19 pandemic and political change on their role.

Intention, But Hybrid: A New Test for Posthuman Agents

When human conduct is carried out through AI systems, the idea of a single intending actor becomes difficult to apply. The paper proposes a court-usable test for attributing criminal responsibility in human-AI systems, centred on who initiates an action, how technology mediates it and whether a meaningful window of human control remains. The aim is to keep responsibility assessable without abandoning the requirement of culpability.

Robot kao pasivni subjekt krivičnog dela

The paper examines what happens in criminal law when a modern robot is physically attacked by a third party, and asks how far existing offences already cover such situations. It sets the question against the wider debate about whether advanced robots could ever hold the status of a subject in law.

Memory as Resistance to Oblivion: The Life and Suffering of Milutin and Draginja Todorović in Social and Family Remembrance

A co-authored contribution on the life and suffering of Milutin and Draginja Todorović, read through the relationship between social and family remembrance.

Pojedini aspekti implementacije savremenih softverskih rešenja u pravosudni sistem i uticaj digitalizacije na pravosuđe i advokaturu

The paper surveys how modern information and communication technologies are entering the judiciary, assessing the potential of e-justice under the current normative framework and the role of artificial intelligence in its further development, including remote hearings in domestic practice.

Pravna pitanja u okviru sistema veštačke inteligencije

Master's thesis on the legal questions arising within artificial-intelligence systems.