Ukraine
Tools
Tasks
Users
Scope
Training
Regulation
Cases
Insights
Information uploaded as at June 2026
AT A GLANCE
Ukraine’s use of AI in criminal justice is expanding but remains fragmented and largely unregulated. Police rely heavily on AI-enhanced surveillance through the nationwide Safe City CCTV network, which includes facial and licence plate recognition and real-time alerts, though concerns persist about security risks and safeguards. Prosecutors use AI tools such as Microsoft technologies and Palantir to document and analyse war crimes, identify suspects, and structure large volumes of evidence, while facial recognition has reportedly been used to locate missing persons and abducted children. Courts employ AI mainly for legal research (e.g. the Supreme Court’s Legal Positions database) and recidivism risk assessment (Cassandra), with broader automation plans disrupted by the war. Defence lawyers increasingly use AI research and prediction tools, and since 2024 judicial and prosecutorial training on digital competence and AI has expanded significantly.
Ukraine has no dedicated statutory framework governing the use of AI in criminal proceedings; its approach instead rests on the strategic Concept of Artificial Intelligence Development, a body of institutional regulations and non-binding soft-law guidance for judges, lawyers, and public servants, and general data protection legislation, with no AI-specific provisions in the Criminal Procedure Code. In parallel, a growing line of Supreme Court and High Anti-Corruption Court case law has established that AI-generated content carries no evidentiary value and cannot substitute for judicial discretion or independent legal judgment, with reliance on unverified AI outputs in court filings treated as an abuse of procedural rights.
Use
As at June 2026, the use of AI in the field of criminal justice in Ukraine is fragmented and unsystematic. In practice, individual tools and systems with elements of AI are already being used or tested by various participants of the criminal justice system (courts, prosecutors, lawyers and law enforcement agencies),to perform auxiliary, analytical, organisational or informational tasks.
Law enforcement
Ukrainian police forces have primarily relied on AI-powered surveillance tools to enhance operations, improving both response times and overall effectiveness.
Predictive analytics
Cassandra is an AI tool used by law enforcement agencies and related services to assess the risk of repeat criminal offences by accused or convicted persons. In particular, probation authorities use Cassandra when preparing pre-trial reports, as well as by social-psychological service workers when working with convicts. An authorised employee enters relevant information about the person, after which the system analyses the inputted data and generates an assessment of the likelihood of reoffending.
Data review and analysis
In Ukraine, pre-trial investigation authorities use Clearview AI, which specialises in facial recognition using images obtained from open sources. Clearview AI therefore assists law enforcement agencies with identifying individuals, using one of the world’s largest databases of facial images, compiled from publicly available resources, including social networks, media, and other web platforms.
Safe City is a nationwide network of AI-enhanced video surveillance of CCTV cameras deployed across Ukrainian cities, managed primarily through CaMaP software platform developed by the Ukrainian company Kobi Software. In larger cities, the system includes real-time AI-powered analytics capable of facial recognition, vehicle licence plate recognition, and vehicle make, model and colour identification. In some cities, the system can proactively generate automated alerts when a specified face or vehicle is detected in the live camera feed. The programme began with early pilot deployments in 2012 in Donetsk. In 2021, police reported that approximately 30 percent of crimes were then investigated with the assistance of Safe City cameras. In 2024, it was reported that there were approximately two cameras per 1,000 citizens. The network has since expanded, both through the installation of new cameras and the integration of existing cameras into the surveillance network.
In 2024, an investigation revealed that Hikvision and Dahua cameras—which account for 74 percent of the CCTV hardware used in Safe City network — pose significant security risks, as Ukrainian authorities fear they are being hacked by Russia for intelligence-gathering purposes as part of its invasion plans for Ukraine. Human Rights Watch has also highlighted the absence of technical and legal safeguards against abuse or current surveillance programmes, pointing to cases of unauthorised and unrestricted access to personal and institutional data by police officers and system operators.
In December 2024, Chernihiv regional police launched an AI-powered video surveillance centre built entirely on domestically developed software, incorporating facial and vehicle recognition capabilities.
In February 2024, Bill No. 11031 was submitted to Parliament (which is still under consideration as at June 2026) to formalise, centralise, and significantly expand video surveillance in Ukraine. The initiative aimed to establish a Unified System of Video Monitoring of Public Security. However, it was withdrawn from consideration on 3 September 2025.
Prosecutors
Evidence review and analysis
The Office of the Prosecutor General has reported using Microsoft tools and Palantir to document war crimes committed by Russian forces on Ukrainian territory. Since February 2022, over 107,000 criminal proceedings related to potential war crimes have been opened. The Office has used AI tools to communicate and receive information from civilians, structure and search files, detect patterns, uncover connections between cases and individuals, identify suspects, pinpoint locations, geolocate crimes and build chronologies. These tools have allowed the Office to quickly and thoroughly document war crime cases.
Ukrainian prosecutors, particularly the Prosecutor's Office for Crimea, have also reportedly used AI facial recognition software, such as Clearview AI, to locate suspects and missing persons. Through this technology, the Office has been able to identify Ukrainian children who had been forcibly transferred from orphanages and shelters, before being adopted by Russian families or sent to so-called ‘re-education’ camps.
Courts
In 2020, Ukraine’s Cabinet of Ministers identified justice as one of the priority sectors for the application of AI solutions. Since then, Ukrainian courts have experimented with various AI tools for legal research and risk-assessment.
Our aim is to give the routine tasks to AI and technology, so as to leave more time for judges to deal with really difficult cases. We understand that some threats exist, but these threats should not stop us from developing and implementing these technologies into the judiciary.
Legal research, analysis and drafting support
In 2021, the Ukrainian Supreme Court launched an AI-powered database called Legal Positions, which includes its own rulings and was developed in response to the Court’s high caseload and the need to make its case law more searchable and practically accessible to judges and other legal users. Its purpose is to make legal research easier and to ensure that the Court’s binding legal conclusions are easier to identify, as a key step towards promoting judicial uniformity. The Supreme Court announced further developments for the database in 2025, now called ‘Database 2.0’, which can flag deviations from established legal positions to help users assess relevance, and quickly extracts key conclusions from court judgments. The project appears to remain in an exploratory, evolving phase, with possible future integration into the electronic court system.
Ukraine’s Unified State Register of Court Decisions holds roughly 120 million court documents and, as at June 2026, is expected to integrate AI in the future.
Decision-making support
In 2021, the Supreme Council of Justice proposed a pilot project to use AI in first-instance courts for administrative offences, such as traffic or parking violations. The reported aim is to develop an AI tool that can assist judges in resolving such cases by identifying inconsistencies in legal interpretations. However, implementation of the tool was interrupted due to the Russian invasion.
Risk-assessment
As noted above, in September 2020, the Ukrainian Ministry of Justice launched Cassandra, an AI-enhanced system that assesses an individual offender’s risk of recidivism for use in both pre-trial and pre-sentence decisions. The system analyses a person’s behaviour using data already collected and available to Ukrainian authorities, together with the individual’s answers to questions asked by a probation or custodial officer. It then synthesises this assessment into a risk score, which judges may consider in their reasoning.
Defence
Legal research, analysis and drafting support
Ukrainian defence lawyers are increasingly drawing on a mix of specialised legal tech platforms and general-purpose generative AI tools. For example, Court on the Palm assesses the similarity of uploaded court documents to previously decided cases and forecasts a likely outcome. Verdictum PRO is a similar tool primarily focused on commercial litigation.
Victims
Under the Criminal Procedure Code of Ukraine, a victim is a direct participant in criminal proceedings and is granted a broad range of rights under criminal procedural law. Pursuant to Article 55(2), a person acquires the rights and obligations of a victim from the moment they submit a statement reporting a criminal offence committed against them or an application to be admitted to the proceedings as a victim. As prescribed by Article 56, the victim’s rights include the right to submit evidence, raise objections and file motions, provide explanations and testimony, review the case materials, and challenge decisions, actions or omissions.
As at June 2026, there are no reported cases of victims using AI in criminal proceedings in Ukraine.
TRAINING
Since 2024, Ukraine’s Professional Standard for Judges has made digital competence (but not specifically AI) a mandatory element of judicial qualification, including the ability to use digital devices; search for, verify, and interpret data; create digital content; comply with cybersecurity principles; communicate in digital environments; and solve problems in those environments.
The National School of Judges of Ukraine has played a central role in training the members of the judiciary. Over the course of 2024 and 2025 the School has conducted a series of specialised seminars on AI and justice. These include programmes such as ‘Artificial intelligence in a judge’s everyday work: between technological support and challenges to professional independence’ (7 December 2025); ‘Electronic Court. Digital transformation of justice in Ukraine’ (23 October 2025); ‘Information and Digital Competence of a Judicial Assistant’ (30 May 2025); ‘Digitalization of Justice and the Use of AI Technologies in Courts’ (6 March 2025); ‘Digital Awareness of a Judicial Assistant’ (29 October 2024), and ‘Digitalization of Justice and the Application of AI Technologies’ (11 June 2024).
For prosecutors, The Training Centre for Prosecutors of Ukraine provides ongoing (though non-mandatory) professional development, particularly in digital evidence processing and data analytics. They have offered programmes such as ‘Artificial intelligence: new opportunities in the work of a civil servant"’(first offered in 3 April 2025, and due to be offered again in March and October 2026) and ‘The use of artificial intelligence technologies in the professional activity of a prosecutor’ (19-20 January 2026).
Finally, the Ukrainian Bar Association has taken meaningful steps to prepare the broader legal community for the growing role of AI in legal practice. Working in partnership with (1) Ukraine's Ministry of Digital Transformation and (2) the Bar Committee on IT, Media, and Data Protection, they organised a voluntary event in Kyiv in September 2025, centred on the launch of a dedicated guide for legal professionals on using AI responsibly.
REGULATION
In Ukraine, there is no dedicated statutory framework governing the use of AI in criminal justice. In the absence of binding legislation, the State's approach is shaped primarily by the Concept of Artificial Intelligence Development in Ukraine, a strategic and programmatic document setting out core principles such as the rule of law, respect for human rights, non-discrimination, transparency, security, and accountability. This is supplemented by internal regulatory acts adopted by individual judicial institutions, together with a body of non-binding soft-law guidance issued by public authorities and professional bodies for practitioners. In parallel, general legislation on data protection and, indirectly, cybersecurity applies to the use of AI systems in the criminal justice context, while a growing body of Supreme Court case law has begun to define the boundaries of permissible AI use in procedural practice, notwithstanding the continued absence of AI-specific rules within the Criminal Procedure Code.
AI regulations
At the national level, there is no specific regulatory framework governing the use of AI in criminal justice. No law or subordinate legislation establishes generally binding rules for the use of AI in criminal proceedings. However, Ukraine has adopted a strategic and conceptual document that shapes the State's approach to the development and use of AI technologies.
Cabinet of Ministers, The Concept of Artificial Intelligence Development in Ukraine (2020)
The Concept of Artificial Intelligence Development in Ukraine was approved by a Resolution of the Cabinet of Ministers of Ukraine (Order No. 1556-r of 2 December 2020), and defines the main directions of Ukrainian State policy in the field of AI development and deployment. Though the document is strategic in nature and does not establish direct legal norms, it does enshrine basic principles that the State must adhere to when using and deploying AI. For example, the Strategy requires the State to:
- Ensure AI systems comply with fundamental rights, democratic values, and the rule of law;
- Adhere to ethical principles in AI design, development, and deployment;
- Ensure AI development and deployment includes appropriate transparency and accountability mechanisms;
- Design AI systems with appropriate safeguards, conduct risk-assessments, and provide transparency;
- Implement AI systems with proper training, monitoring, and human oversight; and
- Facilitate appropriate access to databases for AI development, respecting privacy and security.
Guidelines for Practitioners
Ukraine has developed a set of guidelines for practitioners on the responsible use of AI by lawyers, public servants, and judges. In addition to the below guidelines, the Ukrainian National Bar Association has established a working group to develop regulatory proposals and legislative recommendations on AI, and has convened a professional discussions and training, including a roundtable held in early 2026 on ‘Artificial Intelligence in the Work of Advocates: Ethics, Responsibility, and Legal Process Engineering’, as well as seminars on using language models for drafting and case preparation.
The approaches reflected in the UNESCO Guidelines for the Use of AI Systems in Courts and Tribunals (2025) have been implemented in Ukraine only partially and indirectly. For example, it has engaged with and partially implemented a number of approaches consistent with the Guidelines through strategic documents, soft law recommendations, professional ethics rules, internal regulations of individual courts and emerging judicial practice.
In particular, the Ukrainian framework reflects the principles of: human oversight, the auxiliary nature of AI, personal responsibility for AI-assisted outputs, mandatory verification of results, protection of confidential and personal data, and the inadmissibility of replacing judicial discretion or the assessment of evidence with AI-generated conclusions. These principles are reflected, inter alia, in the Code of Judicial Ethics, the internal AI rules adopted by the Supreme Court and the High Anti-Corruption Court, recommendations for public servants and legal professionals, and relevant Supreme Court case law.
At the same time, Ukraine has not yet established a comprehensive and binding framework covering such aspects emphasized by the Guidelines as algorithmic impact assessments, independent audits, public disclosure of AI systems used by the judiciary, contestability mechanisms and unified mandatory training standards.
Supreme Court of Ukraine, The Regulations on the Use of Artificial Intelligence Technologies by Employees of the Supreme Court Administration (2025)
On 8 December 2025, the Head of the Supreme Court Administration approved the Regulations on the Use of Artificial Intelligence Technologies by Employees of the Supreme Court Administration by Order No. 117. The Regulations establish mandatory internal rules for employees and provide that AI may be used only as an auxiliary tool, without replacing professional judgement, critical assessment or personal responsibility for the final result. Accordingly, AI may not be used to predict judicial decisions, automatically prepare draft judgments or other procedural documents.
Employees must be familiar with the Regulations, undergo periodic training and verify AI-generated outputs against official and primary sources. They must also comply with confidentiality and data-protection requirements. In particular, restricted information may not be uploaded to publicly available AI systems and AI may not be used to process case materials containing personal data.
The Regulations are binding on employees of the Supreme Court Administration. Non-compliance may result in disciplinary or other liability under applicable law, and the use of AI may be subject to monitoring and audit by the Supreme Court Administration.
Ministry of Digital Transformation, Guidance Note on ‘How to Work Safely with AI: Recommendations for Lawyers’ (2025)
The Ministry of Digital Transformation’s Guidance Note on ‘How to Work Safely with AI: Recommendations for Lawyers’ was published in August 2025 and presented through the Ukrainian Bar Association. The Guidance Note functions as a practical risk-mitigation source for legal professionals. According to the document, AI is to be treated as an assistant, not an authority, and advocates remain personally responsible for the accuracy, legality, and ethics of their filings and advice. This includes confidentiality, which requires that lawyers do not enter client-identifying or sensitive case information into open, consumer-facing systems.
Ministry of Digital Transformation, Recommendations on the Responsible Use of Artificial Intelligence for Lawyers (2025)
In July 2025, the Ministry of Digital Transformation adopted Recommendations on the Responsible Use of Artificial Intelligence for Lawyers, addressed to attorneys, prosecutors, judges, and other legal professionals. They emphasise the necessity of professional human oversight, verification of results, limiting the use of AI for analysis and document preparation, and the inadmissibility of replacing a lawyer’s professional judgment with the results of AI.
Moreover, the Recommendations highlight the principle of transparency in using AI, so that lawyers are encouraged to disclose when a document, material or part of it was prepared or generated using AI tools. This approach is considered an important element of a lawyer’s professional integrity and a means of preventing the misleading of other participants in legal relations.
At the same time, these recommendations are purely advisory in nature, and there is no obligation to disclose the use of AI in criminal proceedings at the level of law of procedural rules in Ukraine. The Criminal Procedure Code of Ukraine (see below) does not contain any requirements for mandatory notification of the parties or the court about the use of AI in the preparation of procedural documents or legal positions.
Ministry of Digital Transformation, Guidelines on the Responsible Use of Artificial Intelligence by Public Servants (2025)
In March 2025, the Ministry of Digital Transformation adopted Guidelines on the Responsible Use of Artificial Intelligence by Public Servants. The Guidelines define the principles of legality, non-discrimination, transparency, confidentiality, human control, and subsidiary. The Guidelines recommend using AI only as an auxiliary tool, with full responsibility for the results of AI use resting with the user.
High Anti-Corruption Court, Principles for the Use of Artificial Intelligence Tools in the High Anti-Corruption Court (2024)
The Principles for the Use of Artificial Intelligence Tools in the High Anti-Corruption Court were approved by Order No. 56 of the High Anti-Corruption Court (HACC) dated 19 December 2024. They apply to judges and court staff when using AI in the performance of tasks assigned to the Court, and provide that AI may be used only as an auxiliary tool for organisational, analytical and technical purposes, such as analysing and summarising large datasets, automating repetitive processes, preparing visual materials, creating content, training and professional development.
Users must comply with human rights, professional ethics, the rule of law and information-security requirements. Specifically, official court documents may not be uploaded to AI tools and the use of AI must not affect the independence of the Court or the objectivity of judicial proceedings. The Principles expressly do not apply to the administration of justice itself. The Order also provides for periodic training of judges and court staff on the practical use of AI and AI literacy.
As an internal act approved by an order of the Court, the Principles are binding within the HACC’s internal administrative framework, and responsibility for monitoring compliance is assigned to the Head of the Court Administration. However, the Principles do not establish a separate sanctions regime; any consequences of non-compliance would arise under the applicable disciplinary, employment, ethical or other general legal rules.
When staff use AI, we require them to disclose that to the colleague they hand the work to — so there's a kind of double-check, since both people know AI was involved and mistakes can happen. At the moment there's no equivalent obligation to disclose AI use to external parties, but I think it's a reasonable idea, and we may look at amending our rules to include that in future.
A lot of criminal cases, especially in the sphere of national security or war crimes, are mostly based on electronic evidence — screenshots, videos from websites that may no longer exist, so we cannot always obtain an expert opinion. From one point of view, you cannot let a person who really committed a crime avoid responsibility; but if you only have electronic evidence, there's a threat that it could be fake, and technical expertise cannot always resolve that.
