Russia
Tools
Tasks
Users
Scope
Training
Regulation
Insights
Cases
Information uploaded as at July 2026
AT A GLANCE
Russia has embraced an extensive integration of AI in criminal justice, spanning predictive policing, mass facial recognition surveillance under the "Safe City" programme, biometric monitoring in pre-trial detention centres and corrective colonies, prosecutorial case management through platforms such as AIK "Nadzor WEB", and judicial drafting support that in some instances—most notably at the Vologda Regional Court—has extended to the drafting of judicial decisions themselves. Adoption has been driven substantially by capacity pressures, including chronic shortages of judicial and prosecutorial staff, with the Chairman of the Council of Judges reporting that 35% of Russian courts already use AI in their work. Legal-research tools such as ISKRA and ConsultantPlus's AI assistant are also widely available to defence practitioners, while courts have confirmed that AI assistance in drafting decisions does not itself invalidate a ruling, provided the judge retains responsibility for the final reasoning.
A comprehensive AI law enters into force on 1 September 2026, but it does not directly address AI in criminal proceedings, leaving AI regulation in areas such as state security, policing, counterterrorism, and anti-money laundering to separate future federal laws or Presidential acts. AI use in criminal proceedings is governed by a patchwork of general laws, including the Criminal Procedure Code, the Federal Law on the Police, the Federal Law on Personal Data, and cybersecurity legislation requiring data localisation, none of which was drafted with AI specifically in mind. No finalised professional guidelines on responsible AI use have been issued for judges, prosecutors, or law enforcement, although the Federal Chamber of Advocates has established a working group to develop guidance for advocates.
USE
As at July 2026, AI has been integrated across nearly every stage of Russia’s criminal justice system, from predictive policing and mass facial-recognition surveillance through to prosecutorial case management, judicial drafting support, and legal research tools available to defence practitioners, with adoption particularly advanced in law enforcement and the courts. This expansion has been driven substantially by capacity pressures (including a shortage of judicial and prosecutorial personnel) resulting in AI use that, while officially confined to administrative and supportive functions, has in some instances extended into the drafting of judicial decisions themselves.
Law enforcement
Predictive analytics
The Safe City project (discussed below) includes an ‘Information and Analytical System for Monitoring the Crime Situation’, which provides an interactive map that deploys predictive analysis of crime monitoring and development across the districts of Moscow to determine where offending is more likely.
Data review and analysis
As part of Russia’s ‘Safe City’ project, AI-enabled facial recognition cameras have been deployed widely across urban, transport and road infrastructure in Russian cities, to enable law enforcement officials to identify potential suspects and dissidents. The project uses FindFace Security facial recognition technology developed by Russian vendor NtechLab, which extracts facial features, converts them into biometric templates, and matches them against database records, alerting authorities in the event of a match. The legal identification of individuals remains the responsibility of law enforcement officers in accordance with applicable procedures. Of one million surveillance cameras installed nationwide, approximately one-third are equipped with such facial recognition capabilities. Road cameras also automatically identify vehicle license plates. The number of surveillance cameras across Russia is expected to rise to 5 million by 2030. Research indicates that AI-enabled video surveillance in Moscow has aided the resolution of 70% of Moscow’s criminal cases.
As part of the ‘Safe City’ surveillance system, the Moscow Metro deploys its own, separate, facial recognition system named Sfera, which is used to flag and detain individuals, including journalists.
The Russian Investigative Committee also uses facial recognition software in its investigations, which results in significantly faster identification of wanted individuals. This technology is used to analyse large volumes of data from criminal cases to identify recurring patterns and indicators of criminal behaviour. The final assessment and evaluation of the results in criminal proceedings remain the responsibility of investigators to minimise the risk of bias and errors.
The city of Moscow’s use of facial recognition technology has been at issue in human rights challenges before the European Court of Human Rights. In determining the facts of the case in Glunkhin v. Russia (Application no. 11519/20, 4 July 2023), the Court noted that Moscow’s deployment of live facial recognition technology is highly intrusive on fundamental rights and freedoms under the European Convention of Human Rights.
The Russian Federal Penitentiary Service is piloting the use of AI through its ‘Digital Pre-Trial Detention Centre’ and ‘Digital Corrective Colony’ initiatives. Launched in December 2022 at one of Moscow’s pre-trial detention centres, the digital detention project integrates telecommunications, biometric identification and video analytics to create an Incident Management System, which enables law enforcement officials to be notified of specific risk scenarios. The system relies on over 350 cameras with automated detection of such scenarios, including violence, escape attempts, suicide, fire, and prohibited communication, enabling real-time alerts to staff.
A related pilot was introduced in 2025 at a Corrective Colony (a penal institution in the Russian prison system) in the Vladimir region, expanding these capabilities through voice and facial recognition, creating digital inmate profiles to support identification, monitor movements, and detect suspicious behaviour, with the overall aim of enhancing transparency, safety, and operational efficiency in corrective facilities.
In July 2026, the Ministry of Internal Affairs reported that it was testing AI-based software to detect deepfakes. The testing appears to form part of broader law-enforcement efforts to respond to the growing use of deepfake technologies in online fraud schemes.
In terms of future developments, it has been reported that the Russian government has also increased its investment in the creation of a nationwide surveillance hub operating under GlavNIVTs, the primary Scientific Research Computing centre in Russia, operating under supervision of the Kremlin and in collaboration with the AI Center and the Ministry of the Interior. It has also been reported that Russia is considering the procurement of Russian company NtechLab一which supplies AI for facial recognition of individuals and licence plate number recognition一despite the European Union sanctioning the technology company for the use of its technology in perpetuating human rights infringements against the Russian population.
As part of broader digitalisation schemes of the Russian Interior Ministry, it has been reported that AI would be used by law enforcement to help search for suspects using biological evidence collected at crime scenes.
Prosecutors
AI in the Russian Prosecutor’s Office is being developed as part of a broader digital transformation strategy, formalised in the Order of the Prosecutor General of 11 September 2025 No. 621, which approves a long-term concept for digital transformation through 2030.
Case management
The Russian Prosecutor’s Office uses digital platforms for document processing. For example, AIK ‘Nadzor WEB’ enables fully electronic case management and document workflow, including online registration, routing, and storage of documents. As at July 2026, an AI-enabled system within AIK ‘Nadzor WEB’ is being developed, to automatically process incoming complaints, route them to the competent authority, and notify applicants about the progress and outcome. This system is integrated with the national public services portal and interagency electronic document exchange, allowing communication with nearly 200 public bodies and significantly accelerating response times.
Legal research, analysis and drafting support
It has been reported that AI systems are used within the Russian Prosecutor’s Office to assist with analysing large datasets, forecast crime trends across regions, and assist in reviewing legislation and legal documents, improving the speed and accuracy of prosecutorial work.
Courts
According to the Chairman of the Council of Judges (an elected self-governance body representing judges from all levels and courts of the Russian judiciary) in May 2025, 35% of Russian courts use AI for their work across various regions, most prominently in Moscow. The development of AI use has been attributed to the shortage of judicial staff in Russia. AI is therefore primarily used to take routine procedural work from the judge’s shoulders, and has been deployed alongside broader digitalisation efforts in the judiciary, such as digital case-management platform systems.
Taking into account that a number of subjects have actually formed a model of the so-called 'assembly-line justice', in which the priority is shifted from the substantive assessment of the case to compliance with deadlines and quantitative indicators, the problems of the shortage of qualified judicial personnel are being consistently overcome. . . To this end, the number of justices of the peace increases, the uneven load between judicial stations is redistributed. It is equally important to exclude excessive procedural actions in standard categories of cases, improve order proceedings, develop mechanisms for out-of-court settlement of minor disputes, and review of the jurisdiction of cases.
Case management
Russian courts use AI-supported speech-to-text transcription tools to draft court hearing protocols. In particular, such tools are used in the courts of Bashkortostan, Tatarstan, and the Vladimir and Irkutsk regions. There are also a number of ongoing pilot projects as at July 2026:
- St Petersburg’s courts of general jurisdiction have also been running a pilot since May 2026, as one of ten Russian regions participating in testing a system called UNICA-AI for judicial-system needs, including transcribing hearing audio recordings. The head of St Petersburg's unified court press service noted that audio recognition is near-instantaneous, and that if a basic ‘skeleton’ of the protocol has already been drafted, preparation times fall from around an hour of transcription work down to 15-20 minutes.
- A separate pilot project is reportedly being tested in the Sakhalin region, according to the local Ministry of Public Administration. The tool will prepare a transcript, analyse the audio recording of the meeting, highlight the key points and form a protocol. At the same time, the completeness and correctness of the data will be checked by the secretary and the justice of the peace.
- A similar project was already tested in 2024 at 230 court districts of justices of the peace of the Rostov region. They also used a neural network for speech recognition to automate the preparation of protocols at trials. In the future, the system is proposed to be improved with the help of large language models and to add the function of preparing trial reports from large texts.
Courts in the Amur region use NeuroMagis, which extracts information from the offence reports, fills this information into the court’s online system, and automatically verifies an individual’s identiy. Courts in the same region also use ‘Magis’, a digital assistant that handles case data intake, case registration, and, in some cases, document drafting.
Legal research, analysis and drafting support
In the Komi Republic Supreme Court and Vologda Regional Court, AI supports the preparation of court orders and decisions, by helping judges to search for relevant rules and legislation, analyse documents, explain potential reasons for adhering to or overturning a decision, and compare documents.
The Magis system, described above as being in use by courts in the Amur region, can also be used by judges to draft judicial documents.
In some courts (particularly in Moscow), judges have used AI systems to draft routine procedural documents, such as court orders to initiate or close a case, summons, and sentencing templates for standard offences (eg., traffic crimes, theft). Moreover, in the Vologda Regional Court, it has been reported that AI is used to draft judicial decisions.
Defence
Legal research, analysis and drafting support
Garant, one of Russia’s two leading legal information systems, has launched ISKRA, an AI assistant that can answer legal questions and generate document templates using AI. ISKRA lets a user select a ‘profile’ for more targeted answers一a ‘Universal’ profile, for instance, analyses the entire information bank of the system and produces a structured response一and can generate templates such as employment contracts, powers of attorney, statements of claim, or complaints based on natural-language prompts describing the situation. It can also analyse case law.
ConsultantPlus is Russia’s other dominant legal research system, which launched its own AI-based service in December 2025, intended to help lawyers and other individuals facing legal questions. It answers questions on legal matters, immediately assembles a curated set of relevant documents from the ConsultantPlus database on the topic, and factors in current legislation, caselaw, and expert recommendations when preparing its answer.
Victims
Victims in Russian criminal proceedings have direct legal standing. Under article 42 of the Criminal Procedure Code, a person recognised as a victim is an independent participant in criminal proceedings on the prosecution side and is granted a broad range of procedural rights, including the right to submit evidence, file motions and complaints, participate in investigative actions and court hearings, and appeal procedural decisions.
As at July 2026, there are no reported cases of victims using AI in criminal proceedings.
TRAINING
As at July 2026, there is no mandatory AI training requirement imposed by any authority for judges, prosecutors, law enforcement, or lawyers in Russia.
However, there are voluntary initiatives offered by universities and commercial providers. For example, the Russian State University of Justice (‘RGUP’) (the institution formally responsible under Presidential Decree for judges’ professional retraining and continuing education) is launching a new master’s programme titled ‘Lawyer in the Field of Digital Currencies, Digital Rights and AI Law’ for the 2026/2027 academic year.
The Higher School of Economics offers a voluntary continuing education programme, ‘Practical Applications of AI in Legal Practice’, which is aimed at lawyers, legal department managers, and law students. The programme focuses on the practical deployment of AI tools in legal work, covering areas such as document drafting, contract analysis, prompt engineering, fact-checking, legal research, judicial applications of AI, intellectual property, and risk management.
Similarly, Moscow Digital School offers a voluntary online course, ‘Artificial Intelligence for Practising Lawyers’, designed to help legal professionals integrate AI into their daily work. The course introduces participants to leading AI tools and teaches their use for document review, contract drafting, compliance tasks, legal research, and preparation for court hearings.
REGULATION
Russia adopted its first framework law on AI on 26 July 2026, with most provisions entering into force on 1 September 2026. The law takes a high-level approach and primarily establishes a general regulatory framework for AI development and governance in Russia, while leaving many substantive and sector-specific issues to future regulation. As at July 2026, AI in criminal procedure is governed by pre-existing general laws on criminal procedure, data protection, cybersecurity, policing, and prosecutorial and judicial conduct, none of which was drafted with AI specifically in mind.
AI Regulation
Russia does not yet have a comprehensive legal regime governing the use of AI in criminal proceedings. Instead, the current framework consists of a recently adopted AI framework law, legislation enabling regulatory sandboxes for innovative technologies, and limited judicial guidance. Together, these instruments provide important insight into the direction of Russian AI regulation and its potential implications for the criminal justice system.
Law on Support for AI Development No. 243-FZ, 2026
Russia’s framework AI law was adopted on 26 July 2026 and enters into force on 1 September 2026, with certain provisions taking effect on 1 March 2027. It focuses on “large fundamental AI models”, i.e. AI systems with at least one billion parameters capable of performing a wide range of intellectual tasks.
The law is aimed at supporting domestic AI development and technological sovereignty. It introduces “sovereign” and “national” AI models: sovereign models must be developed by Russian legal entities throughout their lifecycle, while national models may use foreign components. Both must comply with Russian data-localisation requirements, Russian law and “traditional Russian spiritual and moral values”. The Government will determine when only sovereign or national models may be used.
The law also creates a basic framework for labelling AI-generated audio and visual content, by requiring users to be given the ability to identify such content as AI-generated. However, it does not impose a mandatory labelling obligation.
The law does not specifically regulate the use of AI in criminal proceedings. Instead, it provides that separate federal laws or Presidential acts may establish special rules for the use of AI in areas such as state security, operational investigative activities, public order and safety, road safety, anti-terrorism and anti-money laundering.
Federal Law No. 258-FZ, 2020
Federal Law No. 258-FZ of 31 July 2020, ‘On Experimental Legal Regimes in the Sphere of Digital and Technological Innovations in the Russian Federation’, establishes the legal basis for Russia’s AI regulatory sandboxes. The Act defines the goals and principles of experimental legal regimes in the sphere of digital and technological innovation, identifies the circle of participants in such regimes, and regulates relations connected with their establishment and implementation, including their modification, suspension, termination, monitoring, and evaluation of effectiveness and results. The law’s title and core text were amended by Federal Law No. 523-FZ of 28 December 2024, which changed the scope from ‘digital innovations’ to broader ‘digital and technological innovations’, a substantive expansion of the regime’s coverage beyond purely digital or software innovation to technological innovation more broadly.
Law No. 258-FZ establishes the legal mechanism through which a state body could lawfully deploy an AI tool under bespoke rules and exemptions from otherwise-applicable regulation, subject to state oversight and reporting. As experimental legal regimes may be introduced to facilitate innovation in the exercise of governmental powers and functions by public authorities, the framework could potentially be used for AI applications in certain areas of criminal justice administration or criminal proceedings.
Supreme Court Plenum Resolution No. 15, 2026
According to the May 2026 Supreme Court Plenum Resolution on the preparation of civil cases for trial, the parties submitting evidence obtained using AI technologies must disclose this to the court in civil cases. Although this guidance applies to civil proceedings and would be relevant to criminal cases only by analogy, it may indicate the Supreme Court’s emerging approach to AI-assisted evidence and disclosure. Given that the Supreme Court Plenum Resolutions serve as interpretative guidance intended to promote the uniform application of law across Russian courts, the resolution may also provide insight into how similar issues could be approached in other types of proceedings, including criminal cases.
Guidelines for practitioners
As at July 2026, there are no available national guidelines for Russian legal practitioners, judges, or law enforcement on the responsible use of AI in practice. However, the Federal Chamber of Advocates (the national self-governing body for the legal profession of advocates in Russia) has taken active preparatory steps toward developing such guidance for lawyers in general. The FPA’s Vice-President, Sergei Zubkov, has confirmed the Chamber created a working group studying issues connected with lawyers’ use of AI, with a view to developing unified approaches for the profession, while noting in the meantime that AI use by lawyers carries risks for clients, including the risk of privileged information leaking to foreign servers, and that AI outputs must always be double-checked given observed error rates.
In the absence of specific guidance, Russian lawyers are subject to existing rules on professional conduct in Russia:
|
Law enforcement |
Russian police are governed by the Code of Ethics of the Employees of the Internal Affairs Bodies of the Russian Federation. Though it does not specifically regulate AI use, it requires officers to act lawfully, objectively, fairly, and with respect for human rights, dignity, and lawful interests, while avoiding bias in decision-making and ensuring the confidentiality and proportional processing of personal information. These principles suggest that any use of AI by police officers should remain subject to human oversight, comply with legal and ethical standards, avoid discriminatory or arbitrary outcomes, and include safeguards for privacy and data protection. |
|
Prosecutors |
Prosecutors are governed by the Code of Ethics of the Prosecutor of the Russian Federation. The Code requires prosecutors to act lawfully, independently, objectively, impartially and in good faith, to protect individuals’ rights, avoid conflicts of interest, and ensure the confidentiality and security of information obtained in the course of their duties. It also requires prosecutors to exercise independent judgment and to refrain from conduct that could undermine public trust in the prosecution service. These principles suggest that any use of AI by prosecutors would need to remain subject to meaningful human oversight, should not replace the prosecutor’s decision-making responsibility, and must comply with confidentiality, data protection, and fairness requirements. |
|
Judges |
Under the Law on the Status of Judges 1992, judges must comply with the Code of Judicial Ethics. The Code does not expressly address AI, but judges must decide cases independently, based on the facts and law, as well as their own reasoned judgment, without external influence. These requirements limit AI use where it would replace or shape the judge’s ‘inner conviction’ rather than serve as a purely assistive tool under judicial control. |
|
Lawyers |
Lawyers are subject to the Code of Professional Ethics of an Advocate. Although the Code does not expressly regulate AI, its general duties apply to advocates’ use of AI in criminal cases. Advocates must act honestly, competently, and in good faith, protect clients’ interests by lawful means, and maintain respect for the court and other participants. These principles limit AI use where it is careless, misleading, incompetent, or disrespectful (for example, submitting unverified AI-generated material). Moreover, any use of AI by advocates must comply with the strict duty to preserve attorney-client privilege, which covers virtually all information relating to the representation, including the fact of engagement, client communications, legal advice, documents, and case files. As a result, the use of AI tools may raise confidentiality concerns where client information is disclosed to third-party providers, and advocates remain responsible for ensuring that any technology they use does not compromise the protection of privileged information. |
As at July 2026, there is no public evidence that Russia has formally endorsed, incorporated, or otherwise engaged with UNESCO’s Guidelines for the Use of AI Systems in Courts and Tribunals (2025). Russia has, however, otherwise engaged with UNESCO initiatives on general AI ethics policy, including participation in UNESCO-led standard-setting processes through the AI Ethics Committee at the delegation of the Russian Federation to UNESCO.
Talking about replacing a judge with artificial intelligence is, at the very least, premature 一most likely, impossible.
