Use of Artificial Intelligence in Judicial Decision-Making

Background

European judicial systems are entering a period of significant change, marked by the emergence of artificial intelligence as a tool capable, already now or in the near future, of assisting judges in the resolution of a wide range of disputes. Forms of quasi-judicial automated decision-making are already used in private settings, including digital platforms, and in certain non-European legal orders. The development is therefore neither speculative nor remote, but present and expanding.

Against that background, it is not difficult to see why some users of the justice system may regard automated decisions favourably, particularly if they offer greater speed and lower cost than traditional adjudication. Economic and organisational pressures are also likely, over time, to make it increasingly difficult for judges to resist the adoption of AI tools altogether, even where those tools are confined to a supporting role in decision-making.

Any extension of AI to the judicial decision-making process as a whole, or to a substantial part of it, nevertheless raises questions of constitutional and institutional importance. The first concerns the compatibility of machine-generated decisions with the guarantees of the European Convention on Human Rights: whether an automated system could properly be regarded as an ‘independent and impartial tribunal’ for the purposes of Article 6, and whether the systematic use of automated decision-making may sit uneasily with the prohibition of discrimination in Article 14.

A second concern relates to judicial competence. If artificial intelligence were to displace the judge, for example at first instance, there would be a real risk of gradual erosion in the judiciary’s technical and evaluative capacities, with particularly serious implications for appellate review, where a fully developed command of judicial reasoning remains indispensable.

There is, moreover, a human dimension to litigation that cannot be ignored. Judicial proceedings are, by their nature, often burdensome and conflict-laden; even now, litigants may find it difficult to accept an unfavourable decision delivered by a human judge. That difficulty may become more acute if the decision were to come from a machine, perceived as distant, impersonal, and unable to apprehend the human complexity of conflict.

Finally, although the use of AI is likely to expand across a wide range of public and private sectors, judicial decision-making retains a distinctive character. It remains the ultimate safeguard for citizens and businesses alike: the forum in which definitive protection is sought against an alleged wrong. For that reason, the use of artificial intelligence in the administration of justice requires guiding criteria grounded in a clear distinction between uses that are, on the one hand, inevitable and beneficial and, on the other, incompatible with the fundamental values of adjudication. Only on that basis will it be possible to secure broad consensus and preserve space for future developments consistent with human experience and the rule of law.

Aim

The project, in collaboration with the United Nations Educational, Scientific and Cultural Organization (UNESCO), will establish a coherent European framework for the responsible use of AI in legal practice.

Outcomes

• Adaptation of UNESCO’s Guidelines on the Use of AI Systems in Courts and Tribunals for EU Member States;
• Adaptation of UNESCO’s Ethical Impact Assessment Methodology to the EU legal and judicial context;
• Design of a Judicial AI Readiness Self-Assessment Methodology enabling courts to assess their strategic, organisational, and technical preparedness for AI and digital justice initiatives;
• Development of an AI Strategy Toolkit, including templates, checklists, and case examples, to help courts formulate rights-based, EU-aligned AI strategies that can be adapted across Member States and associated countries.