Judicial use of artificial intelligence in Ukraine (Practical Experience, Safeguards and Emerging Challenges)

JUDICIAL USE OF ARTIFICIAL INTELLIGENCE IN UKRAINE

(Practical Experience, Safeguards and Emerging Challenges)

EUROPEAN JUDICIAL TRAINING NETWORK (EJTN)

Ian Bernaziuk, Doctor of Law, Professor JUDICIAL ACADEMY OF THE REPUBLIC OF LATVIA Judge of the Administrative Cassation Court of the Supreme Court (Ukraine) EJTN AI SUMMIT — THE IMPACT OF AI ON THE JUDICIARY (DIGI/2026/01)

Scientific Expert, CEPEJ - GT - CYBERJUST, Council of Europe Riga, 24 – 25 September 2026

AGENDA

1. AI in the Ukrainian judicial environment

2. Ethical and institutional safeguards

3. Recent Ukrainian case law on AI

4. Human verification and professional responsibility

5. Five practical questions for judicial AI use

6. Three red lines

7. Judicial training and meaningful human oversight

8. Key conclusions

Верховний Суд Judicial Use of Artificial Intelligence in Ukraine (Practical Experience, Safeguards and Emerging Challenges)

2

AI in the Ukrainian Judicial Environment Artificial Intelligence Is No Longer a Theoretical Issue Ukrainian courts increasingly encounter AI through submissions prepared by legal professionals, AI assisted research, automated information processing and emerging judicial technologies.

The central question is therefore no longer whether AI will be used in justice, but under what conditions its use remains verifiable, accountable and compatible with judicial independence.

Верховний Суд Judicial Use of Artificial Intelligence in Ukraine (Practical Experience, Safeguards and Emerging Challenges)

3

Ukrainian Ethical Framework AI as an Auxiliary Tool in Judicial Ethics https://zakon.rada.gov.ua/rada/show/n0001415 - 24#Text Article 16 of the Code of Judicial Ethics treats AI as an auxiliary instrument.

Its use must not compromise judicial independence or impartiality, substitute for the assessment of evidence, or transfer final judicial decision making to an algorithm.

Верховний Суд Judicial Use of Artificial Intelligence in Ukraine (Practical Experience, Safeguards and Emerging Challenges)

4

Commentary on the Code of Judicial Ethics Permitted Assistance and Prohibited Substitution https://rsu.gov.ua/uploads/news/risenna rsu no - 14 vid - 02032026 p f631879b7f.pdf The Commentary approved by the Council of Judges of Ukraine clarifies the practical boundaries of AI use.

AI may assist with systematisation, legal research and linguistic work, but it must not determine the outcome of a case, perform the legal assessment of facts or replace effective judicial control.

Верховний Суд Judicial Use of Artificial Intelligence in Ukraine (Practical Experience, Safeguards and Emerging Challenges)

5

Institutional Rules within the Supreme Court Internal Governance Matters as Much as Technology https://court.gov.ua/storage/portal/supreme/rizne/ADMINISTRATIVE_ORDER_HEAD_OFFICE_SC.pdf The Supreme Court has adopted internal rules requiring professional verification of AI generated outputs and restricting the use of public AI systems for protected information.

The institutional approach combines technological assistance with verification, information security and personal professional responsibility.

Верховний Суд Judicial Use of Artificial Intelligence in Ukraine (Practical Experience, Safeguards and Emerging Challenges)

6

Supreme Court Practice: AI Is Not a Source of Law (Commercial Cassation Court), case No. 925/200/22, ruling of 8 February 2024 https://reyestr.court.gov.ua/Review/116984639 A participant attempted to contrast conclusions produced by ChatGPT with the legal position of the Supreme Court. The Court rejected that approach and distinguished authoritative legal sources from machine generated output.

Key point: AI may assist legal work, but its output does not itself acquire the authority of law, precedent or expert knowledge.

Верховний Суд Judicial Use of Artificial Intelligence in Ukraine (Practical Experience, Safeguards and Emerging Challenges)

7

Supreme Court: AI "Hallucinations" (Administrative Cassation Court): case No. 240/14153/24, 15 January 2026, case No. 340/2365/25, 3 February 2026 https:// reyestr.court.gov.ua/Review/133336040, https:// reyestr.court.gov.ua/Review/133764626 The Supreme Court developed this approach in two decisions: case No.

240 / 14153 / 24 (15 January 2026) and case No. 340 / 2365 / 25 (3 February 2026). Both concerned references to non existent or misrepresented judicial decisions and legal conclusions.

Key point: responsibility for AI generated content remains with the person who submits or relies upon it. International recognition: both rulings were included in Damien Charlotin’s AI Hallucination Cases Database, becoming the first entries from Ukraine.

https: //www. damiencharlotin. com/hallucinations Верховний Суд Judicial Use of Artificial Intelligence in Ukraine (Practical Experience, Safeguards and Emerging Challenges)

8

The Practical Lesson from Ukrainian Case Law The Problem Is Not AI — It Is Unverified AI The emerging Ukrainian approach does not prohibit technological assistance as such. It focuses instead on professional verification, good faith and responsibility for the material introduced into judicial proceedings.

AI may accelerate legal work, but speed cannot replace verification against primary legal sources. Верховний Суд Judicial Use of Artificial Intelligence in Ukraine (Practical Experience, Safeguards and Emerging Challenges)

9

Five Practical Questions Traceability and Verifiability

1. Traceability — Where does the information come from?

A judge or lawyer should be able to identify the source underlying an AI generated legal assertion.

2. Verifiability — Can I check it against the original source?

AI output should not replace the authentic judgment, statutory provision, evidence or procedural document. Верховний Суд Judicial Use of Artificial Intelligence in Ukraine (Practical Experience, Safeguards and Emerging Challenges)

10

Five Practical Questions Reason - Giving, Contestability and Responsibility

3. Reason giving — Can I explain the conclusion myself?

A judicial professional must understand and independently justify the conclusion.

4. Contestability — Can the affected person challenge the result?

Legally relevant AI assisted material should not become inaccessible or impossible to contest.

5. Responsibility — Who is responsible if AI is wrong?

Professional responsibility remains human. Верховний Суд Judicial Use of Artificial Intelligence in Ukraine (Practical Experience, Safeguards and Emerging Challenges)

11

First Red Line AI Must Not Replace the Judge in the Assessment of Evidence Evaluation of evidence requires legal judgment, context and responsibility.

AI may assist in organising or retrieving information, but it should not determine credibility, evidentiary weight or the factual conclusion that belongs to the judge.

Верховний Суд Judicial Use of Artificial Intelligence in Ukraine (Practical Experience, Safeguards and Emerging Challenges)

12

Second Red Line AI Must Not Make the Judicial Decision AI should neither make the final decision nor become a practical default that pushes the judge towards a predetermined result.

Human oversight loses its meaning where the judge merely approves decisive reasoning that was not independently assessed and adopted as the judge’s own.

Верховний Суд Judicial Use of Artificial Intelligence in Ukraine (Practical Experience, Safeguards and Emerging Challenges)

13

AI - Generated Judicial Reasoning. Judicial Reasoning Cannot Be Delegated to AI. Kyiv Court of Appeal, case No. 11 кп/824/1818/2025, 30 July 2025 https://reyestr.court.gov.ua/Review/129699665 The appellate court quashed the first instance judgment and ordered a new trial, noting, among other defects, that the reasoning contained theoretical content generated by ChatGPT that was unrelated to the facts of the case.

In doing so, the court expressly referred to the Supreme Court’s ruling in case No. 925 / 200 / 22, illustrating how the Supreme Court’s approach to AI is already influencing lower court practice.

Key point: technological assistance must not substitute for the judge’s own assessment of facts and law. Верховний Суд Judicial Use of Artificial Intelligence in Ukraine (Practical Experience, Safeguards and Emerging Challenges)

14

Third Red Line Confidential Case Information Must Remain Protected Confidential, personal or restricted access judicial information should not be entered into uncontrolled public AI systems.

Responsible AI use therefore depends not only on functionality, but also on cybersecurity, data governance and the availability of secure and institutionally approved technological environments.

Верховний Суд Judicial Use of Artificial Intelligence in Ukraine (Practical Experience, Safeguards and Emerging Challenges)

15

Judicial Training in Ukraine AI Literacy Is Becoming Part of Judicial Competence Ukraine is developing AI related judicial education through the National School of Judges of Ukraine, Council of Europe programmes, the EU Project “Pravo - Justice”, Supreme Court internal training and other professional initiatives.

The central objective is not merely to teach how to use AI, but to develop the ability to verify, challenge and, where necessary, reject its output.

Верховний Суд Judicial Use of Artificial Intelligence in Ukraine (Practical Experience, Safeguards and Emerging Challenges)

16

From Human in the - Loop to Meaningful Human Control Human Presence Is Not Enough A human signature or final approval does not by itself demonstrate meaningful oversight.

Effective human control requires professional competence, access to primary sources, sufficient opportunity for verification and genuine freedom to reject an AI generated recommendation.

Верховний Суд Judicial Use of Artificial Intelligence in Ukraine (Practical Experience, Safeguards and Emerging Challenges)

17

The Emerging Ukrainian Approach Assistance without Transfer of Judicial Responsibility Recent Ukrainian regulation and case law point towards a common principle: AI may strengthen judicial work where it improves access to information, organisation and efficiency, but it must not acquire the authority to determine the judicial outcome.

The decisive safeguards are verification, transparent human reasoning, contestability and continuing human responsibility.

Верховний Суд Judicial Use of Artificial Intelligence in Ukraine (Practical Experience, Safeguards and Emerging Challenges)

18

KEY CONCLUSIONS

AI can support the Ukrainian judiciary, but professional and judicial responsibility cannot be delegated to technology. Recent Ukrainian practice supports five practical safeguards: traceability, verifiability, reason giving, contestability and responsibility.

Three boundaries should remain clear: AI should not replace judicial assessment of evidence, should not become the real decision maker, and should not compromise protected judicial information.

Верховний Суд Judicial Use of Artificial Intelligence in Ukraine (Practical Experience, Safeguards and Emerging Challenges)

19

Selected Previous Publications by the Author

1. Bernaziuk, Ian. Artificial Intelligence in the Ukrainian Judiciary: Charting the Course under the Digital Gavel. 2026.

https: //court. gov. ua/eng/supreme/pres centr/news/ 1891488

2. Bernaziuk, Ian. Artificial Intelligence and the Judicial System of Ukraine: Results of Cooperation in the Past Year.

2025. https: //court. gov. ua/storage/portal/supreme/prezentacii_ 2025 /AI_Ukraine_bernaziuk. pdf

3. Bernaziuk, Ian. Benchmarking Justice: Can AI Uphold the Rule of Law? 2025.

https: //court. gov. ua/storage/portal/supreme/prezentacii_ 2025 / 156 _AI_Benchmarking_Justice_bernaziuk. pdf

4. Bernaziuk, Ian. Artificial Intelligence in Scholarly Research: Personal Scientific Contribution, Academic Integrity

and Responsibility [in Ukrainian]. Zenodo, 2026. https: //doi. org/ 10. 5281 /zenodo. 22257136

5. Bernaziuk, Ian. Artificial Intelligence and Automated Systems in Public Administration and Judicial Proceedings:

Human Oversight, Reason - Giving, Verifiability and Responsibility in the Case Law of the Administrative Cassation Court within the Supreme Court [in Ukrainian]. Zenodo, 2026. https: //doi. org/ 10. 5281 /zenodo. 22274766

6. Bernaziuk, Ian. Sovereign AI: From a Technological Idea to a Matter of State Resilience [in Ukrainian]. 2026.

https: //constitutionalist. com. ua/sovereign ai from a technological idea to a matter of state resilience

7. Bernaziuk, Ian; Barikova, Anna. Discretion in Applying Provisions of Law: Linguistic Prospects for AI and Machine

Learning. ECS Transactions, 2022. https: //iopscience. iop. org/article/ 10. 1149 / 10701. 18545 ecst Верховний Суд Judicial Use of Artificial Intelligence in Ukraine (Practical Experience, Safeguards and Emerging Challenges)

20

Thank you for your attention! The views expressed are those of the author and do not necessarily reflect the official position of the Supreme Court of Ukraine, the Council of Europe (CEPEJ) or the organisers of the event.

21