AEAJ Ljubljana Statement on AI

AEAJ adresses challenges raised for courts by Artificial Intelligence (AI) by its open Statement on AI adopted on 16. May 2025 in Ljubljana.

This statement is the result of a year-long process associating the different components of AEAJ, with the helpful collaboration of IT specialised judges. It presents the professional viewpoint of European administrative judges on issues from the present and near future.

After underlining that AI shall never replace human judges, the statement calls for active public innovation by national judicial systems in order to design appropriate AI tools respecting the fundamental judicial framework. Such tools may help to reshape judicial interactions and improve access to courts, strenghtening citizens' trust into the judiciary and the Rule of law in a decisive period.

The statement is an opening. AEAJ intends to continue working on this decisive topic in the future.

To download the PDF version : AEAJ Ljubljana Statement on AI (English version)

AEAJ LJUBLJANA STATEMENT ON ARTIFICIAL INTELLIGENCE

16. May 2025

When facing AI, justice systems may be tempted by a wait-and-see approach. But passivity brings hidden risks by itself. Without active judicial involvement, Shadow AI - the uncontrolled use of AI - could emerge, undermining transparency and human oversight. Digital divide between lawyers, benefitting from the expansion of legal tech, public authorities expanding AI tools for their own purposes, and administrative judges struggling to keep pace, would impact the balance of administrative justice as a whole. Finally, if judicial actors do not propose changes concerning justice, others will do, and their proposals are likely to be less suited to judges’ needs.

An innovative stance does not mean ignoring risks, but recognising that AI is widespread, that it will impact legal practice, and that it may contribute positively to justice efficiency. AI could be turned into an asset for the rule of law.

1. A Tool for Judicial Efficiency. AI shall never replace human judges, but be seen as an opportunity to enhance, rather than undermine, the administration of justice and the rule of law. Taking into account the increasing complexity of modern adjudication, AI could provide appropriate tools to assist judges in their judicial work.

2. Respecting the Fundamental Judicial Framework. AI must carefully preserve the integrity of the justice system. It should be barred in any case from threatening justice’s underlying principles, such as the right to a fair trial, fundamental human rights, and judicial independence. AEAJ endorses the CEPEJ five principles on the use of artificial intelligence in judicial systems and their environment[1].

3. Transparency and Security. AI systems must be designed, implemented, and used upon a clear and transparent framework. Opaqueness threatens the core values of justice. AI results must be transparent and openly debatable, without unverified delegation to AI systems. Robust measures must be implemented to protect privacy and ensure data security against cyber threats.

4. Redesigning judicial relationships. Justice is fundamentally a human-driven activity. Interactions between the court, the parties and their representatives are essential. AI should not provide a simulacrum of decision-making, but help redesign judicial interactions, facilitate access to justice for everyone and improve the experience of all stakeholders. As AI is based on data provided by society, it has to use it in its service.

5. Professional AI tools for safeguards by design. Therefore, and given the legal, ethical and technological requirements, justice systems should move beyond open-market solutions to engage into co-designing tailor-made AI systems integrating adequate safeguards.

6. Involvement of judges at all steps. In that respect, multidisciplinary teams need to be established where administrative judges would be directly involved into development from the start and actively contribute to training and fine-tuning AI models. Innovative judges, experts and court staff (enablers) must be supported from a human resources perspective. The role of judicial managers to promote, develop and provide efficient tools is essential as well.

7. Training and Education. AI tools require comprehensive interdisciplinary training programs for administrative judges that go beyond simple AI literacy. Administrative judges also need to be informed and trained on technologies in their legal environment, such as AI tools used by public authorities.

8. Role of Judicial Ethics. If the use of shadow AI by a judge or a court staff member may fall under the scope of judicial ethics, this should not be the case for the use of AI tools developed and provided by national justice systems. Judges and court staff as end users cannot be held liable for IT deficiencies.

9. Addressing the use of AI by public authorities. AI may enhance the efficiency of administrative services, though it poses certain risks. Public authorities must indicate during adjudication the use of AI for the issuance of an act and also must be ready to provide the relevant documentation for the courts to decide.

10. Development, innovation and European cooperation. National justice systems should engage into development through public innovation, and invest accordingly. A multi-stakeholder approach should be adopted, engaging judges, legal scholars, AI experts, lawyers and policymakers. Regulatory sandboxes ensure that risks and opportunities are properly assessed in advance.

European States should avoid a silo approach and engage into active collaboration at the European level to share resources, data and technologies.

Recommendation. The AEAJ invites European States and Institutions to adopt national strategies on the development of AI in judicial systems, in the respect of fundamental values, and to invest decisively both into the development of such tools and into the AI literacy of judicial actors.

[1] CEPEJ, European ethical charter on the use of Artificial Intelligence in judicial systems and their environment, 2018