Lviv Guidelines 2026: 22 Practical Guidelines for the Responsible Use of Artificial Intelligence in the Judiciary (Draft for Discussion)

DRAFT — For professional discussion

LVIV GUIDELINES — 2026

22 Practical Guidelines for the Responsible Use of Artificial Intelligence in the Judiciary Preamble Artificial intelligence (AI) can improve access to legal information, the organisation of court operations and the quality of support functions. Its use must, however, preserve judicial independence, personal responsibility, equality of arms, confidentiality and a genuine opportunity to verify automated outputs.

These Guidelines have been prepared with due regard to international standards and recommendations and to the particular features of the Ukrainian justice system, including administrative justice, cyber resilience in wartime, and emerging forms of AI use by parties to judicial proceedings and public administration bodies.

For judges in Ukraine, national legislation and Article 16 of the Ukrainian Code of Judicial Ethics take precedence: the use of AI must not affect judicial independence or impartiality, involve the assessment of evidence or the judicial decision-making process, or contravene the law.

The acceptability of particular use cases under CEPEJ documents does not automatically make them permissible for judges in Ukraine.

For terminological purposes, these Guidelines distinguish generative AI, algorithmic systems that do not use generative models, automated administrative procedures, and agentic systems capable of independently initiating sequences of operations. An administrative act issued automatically is not necessarily the product of AI.

These Guidelines are addressed to judges, court personnel, judicial governance bodies and other entities responsible for deploying technology in the justice system, within the limits of their respective powers and applicable legal restrictions.

I. JUDGES: INDEPENDENCE, INFORMATION SECURITY AND PROFESSIONAL COMPETENCE

1. Verify the legal permissibility of each task

Before using AI, establish the nature of the task, the status of the information concerned, the requirements of procedural law and judicial ethics, and the limits of human oversight.

Practical guidance. Do not automatically extend use cases permissible for general research to the processing of an individual case. Potentially permissible ancillary tasks include translation, technical editing of publicly available texts, and the preparation of educational and organisational materials, subject to applicable legal and ethical restrictions.

Sources: [01, paras. 13, 25–26; 07, Art. 16]

2. Do not delegate judicial powers to AI

The establishment of facts, assessment of evidence, determination of the outcome of a dispute, formulation of reasons and exercise of judicial discretion remain the responsibility of the judge.

Practical guidance. Distinguish technical editing from the substantive drafting of a judicial decision; check compliance with Article 16 of the Code of Judicial Ethics before using a tool. Even ostensibly descriptive AI assistance exceeds a merely technical role if the system determines the legal significance of circumstances, assesses their credibility or formulates arguments material to the determination of the case. Effective judicial control requires the judge independently to determine the outcome and decisive reasons, to have sufficient time and access to primary materials for verification, and to be able to reject AI outputs. Merely approving an automatically generated decision is no substitute for the exercise of judicial powers.

Sources: [01, para. 55; 03, section II(D); 07, Art. 16; 23, Core Principles 2 and 4; 25, para. 3; 29, sections 4.3, 5.2]

3. Preserve the freedom to reject algorithmic recommendations

AI outputs are not binding on judges. The use of an algorithmic recommendation must not create an obligation to follow it or to provide specific reasons for departing from it.

Practical guidance. Distinguish the organisational use of technology from the independent exercise of judicial functions. Ensure that the manner in which AI recommendations are presented, or their automatic display, does not exert de facto pressure on a judge or encourage the judge to treat a system-generated outcome as predetermined.

Sources: [01, paras. 43, 57–58; 03, sections I, II(D); 04; 20]

4. Verify the source, not the persuasiveness of the answer

AI may fabricate a legal provision, judgment or quotation, or inaccurately represent an actual legal holding. Practical guidance. Verify the primary source, date, version in force, factual context and applicability. Another AI system is not an independent primary source. Do not assume algorithmic search results to be exhaustive: a system may overlook material judgments, divergent legal positions and alternative arguments, disproportionately reproduce prevailing approaches, and thereby constrain the development of case law.

Sources: [01; 03, section I; 12; 13]

5. Do not disclose judicial data to unauthorised systems

Non-public case materials, personal data, restricted official information and information concerning judicial deliberations must not be disclosed to services without an appropriate legal and technical basis.

Practical guidance. Do not treat a paid subscription or the mere removal of names as a guarantee of security. Assess re-identification risks, the conditions governing retention of prompts, the possibility of provider access to information, and its use in training models. Use even an institutionally approved system only for specified purposes and authorised categories of data.

Sources: [01, paras. 28–30, 67–68; 03; 08; 12; 25, Recommendations 4, 8–9]

6. Maintain AI literacy and independent professional judgment

Training should cover model errors, undue reliance on automated recommendations, the risk of narrowing the search for alternative legal arguments, data protection, source verification, and the possible erosion of independent legal analytical skills.

Practical guidance. Use practice-based training; assess the ability to identify unreliable outputs, rather than merely proficiency with an interface.

Sources: [01, paras. 42–47; 04, Principle (xiv), paras. 60–61; 11, Arts. 4, 14(4)(b)]

7. Preserve independent legal interpretation and the development of case law

AI must not replace a court in developing new interpretations of the law, deciding whether to depart from established case law, or reviewing the constitutionality of legislation and its compatibility with applicable international conventions, in particular the European Convention on Human Rights.

Practical guidance. When formulating a legal position in an individual case, including when considering a departure from established case law, legal interpretation must be undertaken by the court. This does not preclude permissible ancillary processing of open sources, translation or technical editing, subject to compliance with Article 16 of the Code of Judicial Ethics.

Sources: [01, paras. 25–26; 04; 07]

II. JUDICIAL PROCEEDINGS: TRANSPARENCY, PARTIES’ SUBMISSIONS AND EVIDENCE

8. Establish proportionate limits for documenting AI use

Paragraph 61 of the CEPEJ Guidelines recommends documenting the use of generative AI in the preparation of judicial decisions and ensuring that parties have the right to know which elements of a judicial act were prepared with its assistance. In Ukraine, AI use in judicial decision-making is restricted by Article 16 of the Code of Judicial Ethics.

Practical guidance. Distinguish internal records of permitted technical AI operations from disclosure of AI use to parties to proceedings. Documentation does not render AI use lawful where it conflicts with legislation or Article 16 of the Code of Judicial Ethics. Any obligation to notify parties to proceedings of the use of AI must have an appropriate legal basis and must not compromise the secrecy of judicial deliberations.

Sources: [01, para. 61; 03; 07]

9. Assess participants’ use of AI by its substance and consequences

The court must assess the authenticity of sources, the substance of submissions and compliance with procedural duties, without treating the mere use of AI as evidence of an abuse of procedural rights.

Practical guidance. Paragraph 39 of the CEPEJ Guidelines recommends providing for a procedural duty to disclose the use of generative AI and consequences for non-compliance. In Ukraine, introducing such a specific obligation requires an appropriate legal basis. Irrespective of whether AI was used, courts must verify the authenticity of cited legal authorities, refrain from relying on unverified citations, and determine procedural consequences in accordance with the law and the circumstances of the individual case. Witness testimony and expert opinions require particular care. AI must not replace a witness’s personal perception and account of events or an expert’s independent professional judgment. Where AI has been used materially in preparing an expert opinion, consideration should be given to whether

the nature of that use can be established, the relevant outputs verified, and the expert’s personal responsibility preserved.

Sources: [01, paras. 38–41; 13; 14; 17; 28, paras. 21–31; 30]

10. Protect systems against hidden instructions in case materials

A document submitted by a party may contain hidden instructions intended to alter the behaviour of an AI system processing the file.

Practical guidance. Use only authorised and appropriately secured tools to process parties’ documents. Provide for separation of document content from system-level instructions, access restrictions and verification of processing outputs to prevent execution of hidden instructions embedded in text, metadata or document attachments.

Sources: [01, para. 40; 05]

11. Scrutinise digital materials for manipulation, including deepfakes

AI-generated audio, video and images create additional risks of falsification; a label does not prove falsity, and the absence of a label does not establish authenticity.

Practical guidance. Verify provenance, integrity, metadata and other evidence admissible under procedural law. Where digital material has been collected, transformed or analysed using AI, consider whether the integrity of the original, its processing history, and differences between primary data and automated analysis outputs can be verified. Technical confirmation that a file has not been altered does not, by itself, establish the truth of its content. Do not make the output of an automated detector the sole basis for legal assessment.

Sources: [01, para. 48; 05; 11, Art. 50; 24, section 5.2]

12. Prevent technological inequality between parties

Access to powerful systems and differing technical skills may create asymmetries between parties, particularly where a person is unrepresented.

Practical guidance. Safeguard procedural equality between parties regardless of their access to AI technologies. Take account of the needs of individuals without legal representation or adequate digital skills, and preserve legally available means of participation in proceedings that do not require AI.

Sources: [01; 02; 04, para. 86 and Principle (v); 05; 06]

III. AUTOMATED DECISIONS: ADMINISTRATIVE JUSTICE AND RISKS OF AI USE IN JUDICIAL PROCEEDINGS

13. Review the legality of administrative acts issued automatically

Article 62 of the Law of Ukraine “On Administrative Procedure” permits a fully or partially automated mode only in the cases and according to the procedure prescribed by law; responsibility for the act rests with the administrative authority.

Practical guidance. In disputes, examine the statutory basis for automation, data sources, decision criteria, the ability to identify and rectify errors, the adequacy of reasons and the effectiveness of available remedies. The mere use of an automated mode does not establish that AI was used. Identify the source of any error and the statutory allocation of powers and duties among those who determine the system’s operating rules, ensure data accuracy, operate the system and make legally consequential use of its outputs.

Sources: [09, Art. 62; 05; 06]

14. Take account of the exception to the right to be heard where an administrative act is issued automatically

Article 63(2), point 3, of the Law of Ukraine “On Administrative Procedure” provides that no hearing of the person concerned is conducted where an administrative act is issued in an automated mode.

Practical guidance. When adjudicating a dispute, take account of the statutory exception to the right to be heard, while examining the legality of the automated mode, compliance with other procedural safeguards, the possibility of correcting errors and the effectiveness of subsequent review.

Sources: [09, Art. 63]

15. Require verifiable grounds for algorithmic decisions

Trade secrecy or the complexity of software code must not automatically deprive a person of a genuine opportunity to challenge a decision.

Practical guidance. Within the court’s procedural powers, examine the input data, criteria, methodology and known limitations of an automated system that are material to the dispute. Take account of the data subject’s rights under Article 8(2)(12) and (13) of the Law of Ukraine “On Personal Data Protection”, including the right to know the mechanism of automated processing and the right to protection against an automated decision having legal effects for that person.

In Case C-203/22 (Dun & Bradstreet Austria), the Court of Justice of the European Union, interpreting Article 15(1)(h) GDPR, emphasised the need for an intelligible explanation of the procedures and principles of automated personal-data processing sufficient to enable the relevant outcome to be scrutinised and contested. Effective protection must be reconciled with lawful restrictions on access to information, including the protection of trade secrets and third-party rights.

Sources: [05; 06; 09; 18, p. 7; 27; 31, Art. 8(2)(12)–(13)]

16. Assess the risks of pilot projects in light of the nature of the case and its potential consequences

Standardised proceedings are not necessarily low-risk. In administrative-offence cases, sanctions may include deprivation of the right to drive motor vehicles or administrative arrest, as illustrated by Article 122-4 of the Code of Ukraine on Administrative Offences.

Practical guidance. In evaluating pilot projects, consider not only the standardised nature of proceedings but also the rights at stake and the potential consequences of the decision. Paragraph 36 of the CEPEJ Guidelines envisages pilot projects involving generative AI in specified standardised proceedings. Paragraph 37 requires full human judicial intervention in matters affecting fundamental rights, involving minors or incapacitated persons, or entailing deprivation of liberty measures, limiting generative AI to technical support functions. For Ukrainian judges, those recommendations do not alter the restrictions under Article 16 of the Code of Judicial Ethics: AI may not replace the assessment of

evidence, the exercise of judicial discretion or the judicial decision-making process. Sources: [01, paras. 36–37; 07; 10, Art. 122-4; 11]

IV. JUDICIAL GOVERNANCE: TECHNOLOGY, SECURITY AND ACCOUNTABILITY

17. Ensure cyber resilience and continuity of justice

In wartime, the risks of cyberattacks, network outages, data compromise and dependence on external platforms are particularly significant.

Practical guidance. Assess data hosting, backups, recovery capabilities, access controls, supply chains and the ability to perform judicial functions without AI.

Sources: [01, paras. 29–30; 03, sections II(B)–II(D); 08; 12]

18. Establish requirements for authorised tools and providers

Each product should have defined functions, permitted data categories, access controls, logging policies, retention and deletion arrangements, and a prohibition on unauthorised use of data for additional model training.

Practical guidance. Contracts and technical assessments should provide for audits, incident response, the ability to discontinue a service and control over data processing. At the same time, judges should have access to vetted, secure tools under transparent criteria, without unjustified technological inequalities between courts.

Sources: [01, paras. 29–30; 03, sections II(B)–II(D); 08; 12; 18, p. 8]

19. Involve judges in designing and evaluating technology

Technological change within courts requires the involvement of judges and specialists in law, security, accessibility and data protection, from procurement through to evaluation of results.

Practical guidance. Before deploying a system, define its intended purpose, legal restrictions, anticipated benefits, evaluation criteria and responsible persons. During judicial pilots, test accuracy, security, the effects on judicial independence and procedural rights, and costs. Provide for incident response, restriction or suspension of system use, and renewed risk assessment after material updates.

Sources: [01; 03, sections II(D), II(F); 04; 05; 21; 23, paras. 11–13; 26]

20. Control agentic systems by assessing the cumulative risk of their actions

Agentic AI may sequentially retrieve data, invoke external services, transfer files and initiate operations; risk may also arise from the combination of individually minor actions.

Practical guidance. Assess the cumulative risks of an agent’s sequence of actions. Differentiate permissions to read, modify, delete and transfer information; grant only the minimum access necessary. Require authorisation by a competent person before external transfers of data or other actions with legal consequences. Provide for logging agent actions, the ability to halt them promptly and, where technically feasible, to reverse completed actions or remedy their effects. Do not combine the processing of untrusted external materials with broad agent access to judicial data and services without adequate safeguards.

Sources: [01, paras. 56, 60; 15; 16; 19]

21. Do not turn technical logging into surveillance of judges

AI-use logs, individual performance metrics and algorithmic predictions of an individual judge’s decisions may create risks of interference with judicial independence.

Practical guidance. Distinguish technical security audits from oversight of the substance of judicial activity. Specify the legal basis, purpose, scope of access, retention periods and confidentiality safeguards for logs. Do not permit technical data to be used for unlawful profiling of judges, monitoring of their reasoning, prediction of their decisions, or interference with judicial independence or the secrecy of judicial deliberations. Ensure appropriate control by the judiciary over access to judicial data, the conditions governing their reuse, and the information infrastructure.

Sources: [01; 03; 04, Principle (ii), para. 56; 11, Art. 74(8); 22]

22. Assess AI effectiveness and ensure remedies and appropriate accountability

The effectiveness of AI deployment should be measured by its effects on accuracy, security, workloads, access to justice and human rights—not by the number of licences purchased. Effective remedies must be available where the use of AI infringes rights.

Practical guidance. Provide for impact assessments, independent review, appropriate complaint mechanisms, compensation for harm where statutory conditions are met, and regular anonymised monitoring of the consequences of AI use. Paragraph 76 of the CEPEJ Guidelines recommends that the State assume responsibility for harm caused by judicial errors or failures of generative AI systems, in accordance with existing principles of State liability, and that neither the autonomy nor the technological complexity of such systems exempt the State from its duty to provide redress. In light of monitoring findings, periodically review permitted AI use cases and the corresponding internal rules.

Sources: [01, para. 76; 04; 05; 16]

SOURCES AND MATERIALS CONSULTED

[01] European Commission for the Efficiency of Justice (CEPEJ). Guidelines on the Use of Generative Artificial Intelligence for Courts. CEPEJ(2025)18Final. Adopted at the 45th plenary meeting, 4–5 December 2025.

https://rm.coe.int/cepej-2025-18final-en-draft-guidelines-on-the-use-of-generative-ai-for/48802a4ad1 [02] CEPEJ. European Ethical Charter on the Use of Artificial Intelligence in Judicial Systems and their Environment.

2018. https://www.coe.int/en/web/cepej/cepej-european-ethical-charter-on-the-use-of-artificial-intelligence-ai-in-judicial- systems-and-their-environment

[03] Court of Justice of the European Union (CJEU). Artificial Intelligence Ethics Charter. Adopted on 26 January 2026. https://curia.europa.eu/site/jcms/p1_1000081816 [04] Consultative Council of European Judges. Opinion No. 26 (2023): Moving forward: the use of assistive technology in the judiciary. CCJE(2023)3.

https://rm.coe.int/ccje-opinion-no-26-2023-final/1680adade7 [05] UNESCO. Guidelines for the Use of AI Systems in Courts and Tribunals. 2025.

https://www.unesco.org/en/articles/guidelines-use-ai-systems-courts-and-tribunals [06] Venice Commission. The Updated Rule of Law Checklist. CDL-AD(2025)002. 2025.

https://docs-venice.coe.int/api/Document?pdffile=CDL-AD%282025%29002-e [07] Кодекс суддівської етики, затверджений ХХ черговим з’їздом суддів України 18 вересня 2024 року [Ukrainian Code of Judicial Ethics, approved by the 20th Ordinary Congress of Judges of Ukraine on 18 September 2024] (in Ukrainian).

https://zakon.rada.gov.ua/rada/show/n0001415-24#Text [08] Положення про використання технологій ШІ працівниками Апарату Верховного Суду, наказ від 8 грудня 2025 року № 117 [Regulations on the Use of AI Technologies by Staff of the Supreme Court, approved by Order No.

117 of 8 December 2025] (in Ukrainian). https://court.gov.ua/storage/portal/supreme/rizne/Polozhennya_SHI.pdf [09] Закон України «Про адміністративну процедуру» № 2073-IX [Law of Ukraine “On Administrative Procedure”, No.

2073-IX] (in Ukrainian). https://zakon.rada.gov.ua/laws/show/2073-20#Text [10] Кодекс України про адміністративні правопорушення [Code of Ukraine on Administrative Offences] (in Ukrainian).

https://zakon.rada.gov.ua/laws/show/80731-10#Text [11] Regulation (EU) 2024/1689 laying down harmonised rules on artificial intelligence (AI Act).

https://eur-lex.europa.eu/eli/reg/2024/1689/oj/eng [12] Рекомендації з відповідального використання штучного інтелекту для правників. Україна, липень 2025 року [Recommendations on the Responsible Use of Artificial Intelligence for Legal Professionals. Ukraine, July 2025] (in Ukrainian).

https://storage.thedigital.gov.ua/files/2/72/fcf6c498e0a9f57e7a8521ff9833372d.pdf [13] Supreme Court of Victoria. Practice Note SC Gen 25 — The Use of Artificial Intelligence by Court Users. 2026.

https://www.supremecourt.vic.gov.au/areas/legal-resources/practice-notes/sc-gen-25-the-use-of-artificial- intelligence-by-court-users [14] Касаційний адміністративний суд у складі Верховного Суду [Administrative Cassation Court of the Supreme Court]. Ухвала від 15.01.2026 у справі № 240/14153/24 [Order of 15 January 2026, Case No. 240/14153/24]; ухвала від 03.02.2026 у справі № 340/2365/25 [Order of 3 February 2026, Case No. 340/2365/25] (in Ukrainian).

https://reyestr.court.gov.ua/Review/133336040 https://reyestr.court.gov.ua/Review/133764626 [15] United Nations, Independent International Scientific Panel on AI. Preliminary Report. July 2026.

https://www.un.org/independent-international-scientific-panel-ai/en/preliminary-report [16] United Nations, Independent International Scientific Panel on AI. (2026). AI Agents, Misalignment and the Risk of Losing Human Control: Evidence from the OpenAI-Hugging Face Incident. Thematic Brief, 21 September 2026 (advance unedited version).

https://www.un.org/independent-international-scientific-panel-ai/en/thematic-briefs/ai-agents-misalignment-risks

[17] Civil Justice Council (England and Wales). Use of AI for Preparing Court Documents: Final Report. 29 October 2026. https://www.judiciary.uk/wp-content/uploads/2026/10/Use-of-AI-for-the-Preparation-of-Court-Documents-Final- Report.pdf [18] Webb, P., Raparelli, L., Peterson, D., & Dy, J. (2026). The Right to a Fair Trial in the Age of AI. UNESCO.

CI/DIT/AI/Judges2026/2. https://unesdoc.unesco.org/ark:/48223/pf0000399879 [19] Agencia Española de Protección de Datos (AEPD). Asistentes y agentes de IA: Que la IA actúe por ti, no sin ti.

Decálogo para personas usuarias. 2026. https://www.aepd.es/guias/decalogo-usuarios-asistentes-agentes-ia.pdf [20] Herdağdelen, Amaç; Joren, Hailey; Tamkin, Alex. How AI Agents Shape Experts’ Judgment Calls. Anthropic Research, 27 October 2026.

https://www.anthropic.com/research/agents-judgment-calls [21] Court of Justice of the European Union. Artificial Intelligence Strategy. June 2023.

https://curia.europa.eu/jcms/upload/docs/application/pdf/2023-11/cjeu_ai_strategy.pdf [22] Raad voor de rechtspraak (Netherlands). Advies Uitvoeringswet AI-verordening. Advies 2026/24, 15 July 2026.

https://www.rechtspraak.nl/organisatie-en-contact/organisatie/raad-voor-de- rechtspraak/wetgevingsadvies/wetgevingsadviezen-2026 [23] Courts and Tribunals Judiciary (England and Wales). Position Statement on the Judicial Use of Artificial Intelligence (AI). 9 October 2026.

https://www.judiciary.uk/guidance-and-resources/position-statement-on-the-judicial-use-of-artificial-intelligence-ai [24] Cybercrime Convention Committee (T-CY), Council of Europe. Cybercrime, Electronic Evidence and Artificial Intelligence: Mapping Study. T-CY(2025)1. Adopted by the 35th Plenary on 12 October 2026.

https://www.coe.int/en/web/cybercrime/-/new-mapping-study-on-cybercrime-electronic-evidence-and-artificial- intelligence-approved-during-the-35th-t-cy-plenary [25] Conselho Superior da Magistratura (Portugal). Recomendações no Uso de Inteligência Artificial. Aprovadas pelo Plenário em 8 de abril de 2026.

https://csm.org.pt/wp-content/uploads/2026/04/Parecer-IA.pdf [26] European Commission for the Efficiency of Justice (CEPEJ). 2nd Report on the Use of Artificial Intelligence (AI) in the Judiciary. CEPEJ(2026)7. Adopted at the 46th Plenary Meeting, 18–19 June 2026.

https://www.coe.int/en/web/cepej/latest-documents-adopted [27] Court of Justice of the European Union. Judgment of 27 February 2025, Case C-203/22, Dun & Bradstreet Austria, ECLI:EU:C:2025:117.

https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:62022CJ0203 [28] High Court of Ireland. Practice Direction HC 142: Practice Direction on the Responsible Use of Generative Artificial Intelligence in Court Documents. Issued 29 July 2026; effective 1 September 2026.

https://procedures.courts.ie/practice-directions/full-practice-direction?url=practice-direction-on-the-responsible-use- of-generative-artificial-intelligence-in-court-documents [29] Cour d'appel du Québec, Cour supérieure du Québec, Cour du Québec et Cours municipales du Québec. Lignes directrices sur l'utilisation de l'intelligence artificielle générative par les juges. Adoptées le 31 août 2026.

https://conseildelamagistrature.qc.ca/article/utilisation-de-lintelligence-artificielle-generative-par-les-juges-les- tribunaux-judiciaires-du-quebec-adoptent-des-lignes-directrices-communes [30] Civinini, Maria Giuliana. L'utilizzo dell'IA da parte dell'esperto giudiziario (con particolare riferimento al consulente/perito medico). Questione Giustizia, 3/2026, pp. 287–301.

https://www.questionegiustizia.it/rivista/articolo/l-utilizzo-dell-ia-da-parte-dell-esperto-giudiziario-con-particolare- riferimento-al-consulente-perito-medico [31] Закон України «Про захист персональних даних» від 1 червня 2010 року № 2297-VI [Law of Ukraine “On Personal Data Protection” of 1 June 2010, No. 2297-VI] (in Ukrainian).

https://zakon.rada.gov.ua/laws/show/2297-17#Text

STATUS, DISCUSSION AND FUTURE REVIEW

The “Lviv Guidelines — 2026” are draft recommendations prepared for professional discussion at the international conference to be held in Lviv on 13 November 2026.

This document is not a legislative or regulatory act, an official interpretation of the Ukrainian Code of Judicial Ethics, or an official position of the Supreme Court, the Council of Europe, CEPEJ, the Consultative Council of European Judges (CCJE), UNESCO, the Court of Justice of the European Union or the OSCE. The participation of representatives of those institutions in the discussion does not imply their endorsement of this document.

This draft is open to written comments and proposals. Following the discussion, a revised version may be prepared in Ukrainian and English for dissemination among judges, judicial governance bodies, academic institutions and international partners.

Any endorsement of this document by individual participants or institutions shall be indicated only where confirmed by them and in accordance with applicable procedures.

A further professional review of the Guidelines is envisaged for 2027, taking account of developments in technology, legislation, case law and practical experience of their application.