Journal of Digital Economy, Law and Policy

Calibrating Algorithmic Accountability: Risk-Based Regulation, Enforcement Capacity and Deferral in European Artificial Intelligence Governance

Murager Abylassimov — Mon, 31 Mar 2025 00:00:00 +0500

Risk-based regulation has become the dominant paradigm for governing artificial intelligence, and Regulation (EU) 2024/1689 (the AI Act) is its most complete legislative expression. This article argues that the paradigm's principal weakness lies not in its substantive design but in the mismatch between the obligations it imposes and the institutional capacity available to enforce them. Drawing on doctrinal and comparative-institutional analysis of the AI Act's staged application, the unresolved position of the proposed AI Liability Directive, and parallel developments in the United States, the United Kingdom, China and the Council of Europe, the article identifies three structural pathologies. The first is a deferral dynamic: because risk-based regimes concentrate cost in ex ante conformity assessment, and that cost is borne before any harm materialises, compliance deadlines become politically renegotiable – a pressure already visible in the distance between the August 2026 date for Annex III high-risk obligations and the state of the harmonised standards on which conformity assessment depends. The second is a remedial asymmetry: with the AI Liability Directive stalled, public supervision has no matching private enforcement channel, so deterrence rests almost entirely on administrative capacity. The third is a capacity gradient: the transaction costs of risk-based regulation fall disproportionately on jurisdictions with thin supervisory institutions, which makes wholesale transplantation of the European model inadvisable for many emerging economies. The article proposes a capacity-indexed sequencing model in which transparency duties, procedural rights and sectoral supervision precede general-purpose conformity assessment.

Ex Ante Platform Regulation in the European Union, the United Kingdom and Japan

Zhansaya Zhunissova — Mon, 31 Mar 2025 00:00:00 +0500

Between 2022 and 2025 three major jurisdictions moved digital platform oversight from ex post competition enforcement to ex ante regulation of designated firms. The European Union's Digital Markets Act, the United Kingdom's Digital Markets, Competition and Consumers Act 2024 and Japan's Mobile Software Competition Act share the premise that case-by-case antitrust enforcement is too slow to discipline entrenched gatekeepers, but they differ systematically in how obligations are set, how compliance is determined and how enforcement is escalated. This article develops a four-dimensional institutional framework – designation, obligation-setting, compliance-determination and escalation – and applies it to the three regimes on the public record to January 2025: the designation of seven gatekeepers under the Digital Markets Act, the first year of applicable obligations and the proceedings it produced, the commencement of the United Kingdom's regime and its first investigations, and the phased entry into force of the Japanese statute. Three findings emerge. First, the regimes converge on substantive obligations – anti-steering, self-preferencing, interoperability, data separation – while diverging sharply on procedure. Second, those procedural differences generate a trade-off between speed and legitimacy that each jurisdiction resolves differently. Third, the first compliance cycle suggests that where an obligation is behaviourally simple and verifiable, as in application distribution, it produces observable change quickly; where it is not, as in search and social layers, it produces contestation instead. The article argues that for smaller and middle-income jurisdictions the effective choice lies between building domestic designation capacity and free-riding on compliance changes induced elsewhere.

Sandboxes, Crypto-Asset Rules and Central Bank Digital Currency: A Three-Layer Model of Digital Financial Policy

Kyzzhibek Zamedin — Mon, 31 Mar 2025 00:00:00 +0500

Digital financial policy is usually discussed as a set of separate problems: whether to operate a regulatory sandbox, whether and how to bring crypto-assets within the regulatory perimeter, and whether to issue a central bank digital currency. This article argues that the three are better understood as layers of a single institutional design problem, distinguished by the type of uncertainty each addresses and by the reversibility of the commitment each requires. Sandboxes form an experimentation layer addressing operational uncertainty about novel business models, with low commitment and high reversibility. Codified regimes such as the European Union's Markets in Crypto-Assets Regulation form a rule layer addressing legal uncertainty about asset classification and issuer obligations, with medium commitment. Central bank digital currency forms an infrastructure layer addressing structural uncertainty about the future of the monetary system, with very high commitment and near-irreversibility. Using the innovation-trilemma framework of Brummer and Yadav (2019), the article shows that each layer resolves the trilemma at a different point and that the layers are complements rather than substitutes: a sandbox without a rule layer generates cases that cannot be scaled, and a rule layer without an infrastructure layer regulates private money without providing a public alternative. The argument is developed against the evidence base to January 2025 and applied to the sequencing problem faced by emerging and transition economies. The article concludes that sequencing errors, rather than instrument choice, account for most disappointing outcomes, and proposes readiness conditions for advancing from one layer to the next.