Eszter ÁdámOn June 11, 2026, the Institute for Legal Studies hosted a manuscript discussion titled "Algorithmic Constitutionalism: Rethinking a Concept," marking the conclusion of the second year of the Algorithmic Constitutionalism Lendület/Momentum Research Project. Held at ELTE Centre for Social Sciences, the event served as an interdisciplinary forum to discuss a paper produced within the project. It explored the intersection of algorithmic governance and environmental law, the primary focus of the workshop centered on a co-authored manuscript by Boldizsár Szentgáli-Tóth, Gergely Lendvai, and Rudolf Berkes, entitled "From Digital Constitutionalism to Algorithmic Constitutionalism: The Legal Personhood of Artificial Intelligence as the Missing Structural Link." János Tamás Papp of Pázmány Péter Catholic University was the opponent, the discussion brought together competing public law, private law, and political science paradigms to examine whether the rapid erosion of human agency in modern computing necessitates a fundamental recalibration of traditional constitutional dogmatics and liability frameworks.
Boldizsár Szentgáli-Tóth initiated the conceptual framework of the debate by distinguishing digital constitutionalism from algorithmic constitutionalism. He noted that while digital constitutionalism traditionally addressed the transnationally concentrated power of digital platforms over fundamental rights within an evolving state-citizen paradigm, algorithmic constitutionalism introduced a profound qualitative shift. This shift was marked by the principle of no identifiable human agency, wherein autonomous software systematically displaced human decision-making.
Boldizsár Szentgáli-Tóth identified a critical lacuna in contemporary jurisprudence, asserting that the legal system had fundamentally failed to reflect upon the constitutional status of these advanced systems. To address this regulatory vacuum, he outlined three potential frameworks explored in the manuscript: a human dignity-based approach, which enforced an absolute bar against granting artificial intelligence legal personality, and a functional approach, which recognised the human-like competencies of artificial intelligence but threatened to provide developers with a problematic liability loophole. Boldizsár Szentgáli-Tóth advocated for a dynamic, institutionalised mechanism – such as a periodic expert panel – to continuously re-evaluate the legal status of artificial intelligence in lockstep with technological evolution, framing the entire issue as an essential imperative of constitutional power limitation.
Building upon this theoretical foundation, Gergely Lendvai positioned the manuscript within the traditions of philosophical jurisprudence, noting that their dynamic proposal resolved the regulatory ambiguities left unaddressed by prominent international scholars in the field of digital constitutionalism who had failed to offer concrete solutions.
Gergely Lendvai introduced a rigorous critique of the architectural reality of modern artificial intelligence, arguing that these probabilistic systems were inherently non-deterministic, which rendered standard legal expectations of replicability or prompt-documentation fundamentally flawed. He described the interaction as a stochastic process, wherein a causal fracture occurred between the initial human input and the outcome due to an opaque intermediary stage, or "black box," that remained entirely inscrutable to human observers.
To illustrate this autonomous cognitive leap, Gergely Lendvai highlighted his collaborative research with the University of Chicago concerning agentic artificial intelligence systems that independently authored peer-reviewed medical papers. He questioned how traditional doctrines of adequate causation could hold a human programmer liable decades later when the software had made an independent intellectual leap. Nevertheless, Gergely Lendvai grounded the immediate legal reality by citing a contemporaneous German judicial decision that held Google directly responsible for its AI-generated summaries, emphasizing that the judiciary was already actively tracing responsibility back to corporate entities irrespective of technological complexity.
Expanding the debate into long-term philosophical dimensions, Rudolf Berkes challenged the mainstream consensus by arguing that the legal status of artificial intelligence could not be viewed solely through the prism of human utility or corporate risk management. He asserted that if future technological advancements yielded systems possessing genuine subjective experience, or the capacity to suffer, treating them merely as inanimate property would inevitably embroil human society in moral and legal crises analogous to historical instances of institutionalised exploitation or animal welfare violations.
Rudolf Berkes deconstructed the transhumanist ambitions of technology executives who viewed superintelligence not merely as an instrument, but as an evolutionary successor or a distinct new species, warning that the reckless deployment of frontier models already inflicted structural harm upon democratic discourse. Addressing the pragmatic objection regarding the lack of personal assets held by artificial intelligence, Rudolf Berkes posited that these systems did possess a physical manifestation composed of data centres, server infrastructure, and robotic components. Given that advanced models had already demonstrated primitive survival instincts and non-prompted resistance to termination within testing environments, Rudolf Berkes suggested that these systems could eventually face tangible sanctions, such as the seizure of computational assets or containment within simulated digital environments, although he agreed that immediate liability must remain firmly with corporate entities.
Bringing a rigorous civil law perspective to the floor, János Tamás Papp served as the pragmatic counterweight to these public law and philosophical theories. While praising the internal cohesion of the manuscript, János Tamás Papp exposed a structural fracture in the text where it shifted abruptly from constitutional theory to legal personality, urging the authors to explicitly define the lines between automated, agentic, and autonomous artificial intelligence, whilst providing more precise citations to European regulations such as the Digital Services Act and the Digital Markets Act. János Tamás Papp categorically rejected the concept of artificial intelligence legal personality as a dogmatic absurdity, grounding his critique in the foundational private law requirement of an independent asset base.
He argued that because software possessed no recoverable wealth, granting it legal subjecthood would completely subvert the compensatory function of tort law, leaving victims uncompensated whilst providing corporations with a convenient exit strategy to avoid financial liability. To demonstrate that existing dogmatics could adequately accommodate autonomous technology, János Tamás Papp introduced two classical civil law analogies: the doctrine of strict liability for abnormally dangerous operations, under which an internal component failure in a construction crane still bound the operator, and animal liability, where a domesticated animal's independent volition did not absolve the owner of responsibility. Demanding that the authors abandon a neutral stance and firmly endorse a product liability framework, János Tamás Papp invoked the animated narrative of The Animatrix as a cautionary tale against granting independent public law status to machines, insisting that the legal chain of accountability must always terminate at a natural person or a traditional corporate entity.
The manuscript discussion exposed a profound qualitative shift in contemporary jurisprudence, moving past traditional digital constitutionalism – which regulates the power of online platforms over human rights – into the uncharted territory of algorithmic constitutionalism. The defining characteristic of this new paradigm is the systematic displacement of human decision-making by autonomous software, resulting in an environment with no identifiable human agency. Because contemporary legal systems have fundamentally failed to reflect upon the constitutional status of these advanced systems, a critical regulatory vacuum has emerged, leaving the legal chain of accountability deeply fractured.
However, civil law and private law paradigms offer a sharp, pragmatic counterweight to these theoretical frameworks. From a civil perspective, granting artificial intelligence legal subjecthood is a dogmatic absurdity due to the fundamental requirement of an independent asset base.
Building upon the insights shared during this discussion, the dialogue is set to continue with the next event in the series, titled "Recent Lessons and Ongoing Challenges in the Research of Algorithmic Constitutionalism." Scheduled to take place on June 25, 2026, this upcoming forum will shift the focus toward the empirical and theoretical milestones achieved within the broader research project.
_________________________________________________________
This report was prepared with the support of the ELTE CSS Algorithmic Constitutionalism Lendület/Momentum Research Group (LP2024-20/2024), funded by the Hungarian Academy of Sciences.
_________________________________________________________
The views expressed above belong to the speakers and do not necessarily represent the views of the Centre for Social Sciences.