DE
DE

Artificial Intelligence
Artificial Intelligence and Justice

Ensuring responsible use of AI by courts and judges
2.5 Minutes
AI judge

Artificial Intelligence (AI) is no longer a future prospect for justice systems – it is a present reality. Judges already use large language models to draft procedural documents; courts deploy algorithmic tools to manage caseloads; risk-assessment instruments inform pre-trial detention decisions; and automated translation expands access to multilingual justice. Much of this is happening in the absence of institutional policies, regulatory frameworks, and thus adequate guardrails.


The question that policymakers and court leadership face is therefore not whether to use AI in the justice sector, but how to govern its use in ways that strengthen, rather than undermine, the rule of law and the fundamental rights of those who encounter it. The stakes are high: courts are the last institutional guarantor of rights, the site where law meets individuals at moments of maximum vulnerability. Failures of AI in this domain are not merely technical malfunctions – they can easily become failures of justice.


AI in the justice sector offers opportunities – in efficiency, in access to justice, and in the quality of legal information – but realizing them depends on governance conditions that are, in most jurisdictions, not yet in place. Three categories of risk demand primary attention: threats to fundamental rights (particularly the right to a fair trial, non-discrimination, and the right to an effective remedy); threats to judicial independence; and systemic risks arising from the concentration of AI capacity in a small number of states and private companies.


The present analysis points to three key recommendations. First, every court and judicial administration should adopt a formal policy on AI use before deploying any tool, covering permitted and prohibited use cases, data governance requirements, transparency obligations, oversight mechanisms, and access to remedies. Second, deployment should proceed in phases – beginning with lower-risk administrative functions – and be accompanied by mandatory fundamental rights impact assessments. Third, the human judge must remain at the center of judicial decision-making: AI outputs shall never constitute binding determinations, and the motivation of every judicial decision must be the judge’s own intellectual product.

AI and Justice

AI and Justice

AI is already present in justice systems worldwide. The question is no longer whether to use it, but how. Used well - with the right tools and under the right oversight - AI can make justice systems faster, more accessible, and better resourced. Used carelessly, it can corrupt the very thing it is supposed to support. This paper highlights the complexity and key dimensions of AI and Justice” and gives concrete guidance on how to ensure the responsible and human rights centered use of AI by courts and judges.

More

About the Author

Zoltán Turbék

is an independent expert specialized in AI governance/regulation, multilateral diplomacy and international law. He is a former career diplomat and UN (IAEA and WHO) legal officer with over 20 years of experience in international law and multilateral policy-making processes. He actively contributed to drafting the Council of Europe’s AI treaty and the UNESCO Recommendation on the Ethics of AI, and has recently been elected to serve as a member of the International Law Association’s new AI & Tech Law Committee.