Kazakhstan

Information uploaded as at July 2026

AT A GLANCE

Kazakhstan is actively adopting AI into its criminal legal system. The strongest public evidence of operational deployment is on the policing and prosecutorial side, where state-built AI systems are integrated into official investigative and analytical platforms. Courts in Kazakhstan have also begun deploying AI tools, particularly in civil and administrative proceedings, though public records do not yet indicate similar adoption in criminal proceedings. Defence counsel and victims do not yet have state-provided AI tools integrated into criminal proceedings, though bar organisations are training lawyers on commercial generative AI and a public-facing AI legal assistant is available to citizens.

Kazakhstan’s regulatory framework for the use of AI in the criminal legal system is still emerging. As at July 2026, there do not appear to be any statutes or binding criminal procedure rules that specifically authorise, restrict, or prescribe safeguards for AI use in investigations, prosecution, or adjudication. Instead, any legal constraints on AI in criminal cases would likely be applied indirectly through pre-existing legislation concerning the regulation of personal data, and the rights enshrined in the Kazakh Constitution. Similarly, in the absence of mandatory or systematic training in relation to AI use for Kazakh legal professionals, the the General Prosecutor’s Office, the judiciary, and various bar associations have begun providing AI-focused training to their employees and members.

USE

Law enforcement

Kazakhstan’s law-enforcement AI landscape is highly developed, consisting mainly of data review and analysis tools, with some operational support tools.

Operational support

In June 2025, Kazakhstan’s General Prosecutor’s Office and Ministry of Internal Affairs jointly introduced a ‘Digital Assistant to the Investigator’. The Digital Assistant has been integrated into the Unified Register of Pre-Trial Investigations, the country’s central criminal case tracking system. Official descriptions describe three functions of the tool:

1. Investigation Planning helps police officers classify cases, propose investigative approaches, identify relevant witnesses and experts, and incorporate judicial precedents into the investigator’s working materials.
2. Testimony Processing transcribes interviews in real time and analyses speech patterns and behavioural signals to flag possible inconsistencies.
3. Document Generation prepares draft investigative documents, including final procedural documents that draw on relevant judicial precedents, for the investigator’s review and signature.

Predictive analytics

A Senate Committee Session in May 2025 referred to the use of ‘predictive analytics’ by the Ministry of Internal Affairs (MIA), alongside intelligent data analysis, image recognition, and decision support. In July 2025, the Committee on Legal Statistics and Special Records separately described its ‘digital assistant’, reportedly capable of finding connections between cases, building forecasts, and generating recommendations for both investigation and crime prevention.

Data review and analysis

The Ministry of Internal Affairs, in cooperation with the National Security Committee and the Ministry of Digital Development, operates a National Video Monitoring System, also known as the ‘Safe City’ Platform. The system uses AI-based computer vision to record offences automatically, perform facial recognition, identify vehicles by make, model, and colour, detect unattended objects, and analyse behaviour patterns in public spaces. Data is gathered from Kazakhstan’s extensive network of CCTV surveillance cameras—as of late 2024, there were approximately 1.36-1.4 million surveillance cameras nationwide, at least 19,000 of which had AI-specific capabilities—and transmitted in real-time to Operational Management Centres staffed by police officers. The Safe City system reportedly enabled the identification and detention of more than 500 wanted persons between June 2024 and June 2025. During the 2024 World Nomad Games in Astana, the system also helped to locate missing persons and supported the arrest of 76 wanted individuals.

To maximise the operability of the Safe City system, the President of Kazakhstan signed an amendment to the Law on the Prevention of Offences, expressly prohibiting wearing clothing in public places that interferes with facial recognition. Exceptions were introduced for medical and professional purposes, but not for religious purposes. Civil society organisations have expressed concern about the human rights implications of the Safe City system, including the right to privacy and freedom of religion.

Sergek is an AI-enabled surveillance system designed primarily for traffic management that has not yet been integrated into the Safe City platform. It has operated in Astana since 2017 and was later introduced in Almaty and Shymkent. The platform relies on AI-based computer vision to monitor road conditions, identify traffic offences, and generate analytical information for the police. During the COVID-19 pandemic, it was used to assess whether drivers had departed from their usual routes in breach of movement restrictions.

Qorgau is another AI traffic enforcement system deployed by the Ministry of Internal Affairs. It uses computer vision to detect traffic violations from street cameras, such as speeding, running red lights, and improper lane usage. Alerts are sent to nearby police officers, allowing them to respond quickly. Those officers are equipped with approximately 21,000 AI-enabled tablets, and also carry smart badges, which go beyond conventional body cameras by providing continuous tamper-proof video recording, GPS tracking, encrypted data handling, and emergency alert functions. Plans are also underway to add facial recognition capabilities and integrate the smart badges directly with law enforcement databases.

Kazakhstan’s law enforcement authorities have also introduced AI-powered video surveillance across the country’s correctional facilities. Following a presidential mandate issued in 2020, the Ministry of Internal Affairs installed more than 39,500 high-definition cameras with AI analytics across 78 prisons. These capabilities include facial recognition, event detection, and licence-plate identification. The system transmits footage continuously to a centralised Situation Centre, and further integration of more advanced object- and event-recognition algorithms is planned. According to official claims, in the first half of 2024 the system helped to prevent 62 suicides, six escape attempts, 27 violent conflicts, and identify 32,000 violations of detention protocols.

Law enforcement agencies in Kazakhstan also use Cellebrite UFED—a mobile forensic extraction tool that relies on AI and is capable of retrieving messages, photos, browser histories, and deleted data—in criminal investigations.

Civil society organisations have raised human rights concerns over law enforcement agencies in Kazakhstan using Cellebrite UFED, arguing that the software is being used to unlock phones and spy on and target activists and journalists without meaningful consent, transparency, or effective judicial oversight.

Tools are also being developed for AI-assisted data analysis in criminal investigations at local levels. For example, the Kyzylorda Regional Prosecutor’s Office has developed an IT TERGEU tool, which uses AI to review data and identify criminal incidents or missing persons. As part of the project, investigative teams were able to check 689 mobile phone numbers in connection with 496 criminal cases. The checks revealed 134 criminal offences causing damages totalling more than 200 million tenge, and 62 criminals were apprehended. In addition, 51 people who had been reported missing were located within a single day.

Prosecutors

Most AI tools deployed by Kazakhstan’s General Prosecutor’s Office as at July 2026 are designed to help prosecutors and investigators process large volumes of data, identify risks and inconsistencies, detect criminal networks or illegal markets, trace financial flows, generate procedural documents, and move cases more efficiently through the investigative stage.

Legal research, analysis and drafting support

The General Prosecutor’s Office has publicly stated that prosecutors are using AI in ‘forecasting and analytical activities.’ In June 2025, General Prosecutor Berik Asylov asserted that ‘[a]ll digital technologies are oriented toward the prevention of offenses and the early forecasting of potentially dangerous phenomena,’ including the Integrated Security Information Resource Complex, which searches for digital traces of wanted individuals across various state agency databases and surveillance cameras, and identifies them for detention.

In addition, the Ministry of Justice is reportedly developing E-Saraptama 2.0, an upgraded forensic examination platform that will integrate AI into the assessment of bodily injury cases and the preparation of expert conclusions. The authorities plan to expand the tool’s use to other types of forensic examination. According to the Deputy Head of the Committee on Forensic Examinations, the use of AI is expected to make forensic conclusions more objective and reduce corruption risks, particularly as such conclusions may serve as evidence in criminal and civil proceedings.

Evidence review and analysis

Digital Supervision is an AI-enabled information-analytical platform of Kazakhstan’s General Prosecutor’s Office. It integrates data from state databases, public cameras, drones, and other systems to support prosecutorial oversight, public-safety monitoring, the search for wanted persons, probation control, and preventative law enforcement analytics. It was announced at the General Prosecutor’s Office Coordination Council in October 2025, and has since reportedly contributed to a 24% reduction in wanted persons.

The General Prosecutor's Office deployed ProkAi, an AI tool for detecting illegal drug marketplaces. The tool operates continuously to detect online drug trading. The tool searches the internet for drug-sale activity at a much larger scale than a human investigator could. It also helps to identify dropper accounts (bank accounts used by intermediaries to receive, move, or launder proceeds from drug sales). As at July 2026, it has reportedly identified 12,000 accounts linked to online drug trading.

Courts

As at July 2026, the courts of Kazakhstan use several AI tools. Some are simple, designed for specific purposes, and integrated into pre-existing technological services. Others, however, are more advanced and offer a wide range of services, including tools designed to support judicial decision-making.

Case management

Automation has been applied to the selection of juror candidates in criminal cases. In late 2024, it was reported that AI-prepared lists of juror candidates had been prepared in 237 criminal cases.

E-Soty AI is an AI ‘smart filter’ for statements of claim, designed to identify formal filing defects before they lead to returns or re-filing. E-Soty AI can flag incorrect jurisdiction, missing documents, and arithmetic errors. A 2025 report claims that the filter reviewed more than 500,000 claims in the first nine months of 2025 and identified errors in 15% of them.

As at July 2026, Kazakhstan courts are testing a courtroom audio-visual system with facial recognition and real time speech-to-text conversion.

Legal research, analysis and drafting support

‘Robotic assistants’ prepare draft rulings in routine, uncontested matters such as travel restrictions, which judges can review and sign. Such assistants can handle as much as a quarter of a judge’s workload, contributing to a reduction in case backlogs, and leading to faster case resolution times.

The Digital Analytics of Judicial Practice Service, used by the judiciary since 2022, compares judicial decisions, identifies anomalies, and predicts outcomes. The system allows judges to understand legal precedents that have been relied upon in similar cases, though judges retain final decision-making authority. As at July 2026, the platform appears to be fully operational as part of Kazakhstan’s digital judicial infrastructure, and is exclusively reserved for judges and other court staff.

Decision-making support

The Digital Judicial Analytics platform (see above) can prepare draft rulings in routine or uncontested cases where the applicable law is clear, mostly in civil cases, though the judiciary is exploring other types of cases to which the system’s automated features could be expanded. The responsibility for judicial decisions remains with human judges, who are tasked with reviewing the suggested draft.

An Assistance to Judges tool reportedly suggests what punishment would be ‘optimal’ in criminal cases in Kazakhstan and calculates the ‘accuracy and fairness of sentences’. It has been reported that the system’s use has led to a 20% decrease in prosecutorial appeals against sentences between 2020 and 2024.

Defence

As at July 2026, there is a variety of tools designed to support lawyers in judicial proceedings, from both the public and private sectors in Kazakhstan.

Administrative support

The Digital Bailiff is a Ministry of Justice AI system launched in August 2024 as part of the country’s wider digitalisation and automation of justice services. The goal of the tool is to replace bailiffs, who are authorised enforcement officers that make sure court orders or other official enforcement documents are carried out, for example by collecting unpaid fines or debts, originating in civil, commercial, or criminal law rulings. The tool automates simplified enforcement proceedings for relatively small debts (including administrative fines and tax debts) by sending citizens SMS or eGov Mobile and push notifications, enabling online payment, and automatically lifting enforcement measures such as account arrests once the debt is paid. If the debt is not paid within the simplified procedure period, the case can be transferred to a human bailiff under the ordinary process. As at July 2026, the Digital Bailiff is operational with nationwide implementation reported by the Ministry of Justice.

Legal research, analysis and drafting support

The Ministry of Justice and National Information Technologies JSC introduced E-zan Komegi (or ‘E-legal assistance’) in 2020, which functions as Kazakhstan’s Unified Information System for Legal Assistance. The tool consists of a website and mobile application, and allows citizens to receive online or video consultations from lawyers, access free State-guaranteed legal assistance, and ask legal questions through an AI chatbot. Lawyers and legal advisers accessing the system can check information on individuals and organisations, gather evidence more efficiently, exchange requests with State and non-State bodies, and maintain registries. The system is reported to use AI capabilities to forecast case outcomes and assist forensic experts by preparing materials for human verification.

PRG.KOMEK and AdiletAI are privately developed legal technology tools in Kazakhstan that use AI to support legal research and legal practice. Both tools are designed to help lawyers and other users work with legal information more efficiently, by allowing users to ask questions in ordinary language, search through relevant legal materials, receive structured answers, and access references to legal sources.

Victims

As at July 2026, the use of AI tools by victims is a developing area of AI deployment in Kazakhstan, and still in the early stages.

Legal research, analysis and drafting support

In July 2025, Karaganda University and IT company ItByDesign launched LegalExpert.kz, described as the country’s first free legal assistant based on AI. The service covers criminal, administrative, and civil questions in Kazakh and Russian, providing step-by-step guidance, links to legislation, examples of court decisions, and forecasts of possible case outcomes. A pilot of the service was launched in Karaganda Oblast, but it is also open to all residents of Kazakhstan.

As discussed above, the Ministry of Justice and National Information Technologies JSC have introduced E-zan Komegi (or ‘E-legal assistance’) in 2020, which functions as Kazakhstan’s Unified Information System for Legal Assistance. The tool consists of a website and mobile application, and allows citizens to receive online or video consultations from lawyers, access free State-guaranteed legal assistance, and ask legal questions through an AI chatbot.

TRAINING

As at July 2026, there is no systematic or mandatory training for law enforcement, judges, prosecutors, or lawyers in Kazakhstan. Instead, AI-related training in the legal and justice sector has developed gradually and unevenly, through a combination of conferences, professional-development courses, technical workshops, and broader digital-skills initiatives. These efforts differ in scope and ambition.

One of the earliest and most visible initiatives took place in May 2023, when the Law Enforcement Academy, together with the Supreme Court and the Union of Judges of the Republic of Kazakhstan, organised an international conference titled ‘Artificial Intelligence and Big Data in the Judiciary and Law Enforcement: Realities and Needs.’ The Law Enforcement Academy is a special-status educational organisation established in 2015. Operating under the Prosecutor General’s Office, it provides initial professional training for entrants to the law-enforcement service, among other government services. The event was supported by UNESCO and the Organization for Security and Co-operation in Europe, and brought together more than 100 participants. Although this conference was closer to a knowledge-sharing event than a training course, it helped place AI, cybersecurity, and the use of AI in courts and law enforcement on the agenda of criminal justice actors in Kazakhstan.

For law enforcement officers and prosecutors specifically, the Law Enforcement Academy has offered several sessions on AI and related technologies between 2025 and 2026. Moreover, in February 2026, the UN Office on Drugs and Crime, Kazakhstan’s Financial Monitoring Agency and the Prosecutor General’s Office organised a simulation training in the Astana region on AI and open-source intelligence for financial investigations and asset recovery, aimed at, inter alia, law enforcement officials, prosecutors and representatives of the Supreme Court.

Judges have also received training, although this has often been framed as digital justice, digital evidence, and judicial modernisation rather than as standalone AI training. For example, in March 2026, the Academy of Justice under the High Judicial Council ran an online professional-development course for judges of criminal chambers. These courses included sessions on IT-related crimes, blockchain analytics, digital evidence, open-source intelligent tools, and digital ethics in the context of the transformation of the judicial system.

Lawyers have also begun to receive practical AI training through bar associations. The Almaty City Bar Association offered a webinar on the foundations, risks, and responsibility of AI for advocates, trainee advocates, and legal consultants in January 2026. The Almaty City Bar Association’s 2025 training report also indicates that there were sessions on using ChatGPT-4 effectively in an advocate’s work, applying AI in legal activity through ChatGPT, and understanding the role of AI in jurisprudence and the automation of legal processes. Similarly, the Astana City Bar Association listed short professional-development sessions in 2025 on ‘AI in the Activity of the Lawyer’ and ‘Practical Examples of Using AI in Legal Practice’.

Finally, academic institutions have contributed through legal conferences and university-based discussions. Al-Farabi Kazakh National University, for example, organised a 2026 event on AI and digital technologies in civil procedure, with participation from legal scholars, practising lawyers, judges, and advocates.

REGULATION

As at July 2026, there are no statutory rules specifically governing the use of AI in criminal proceedings in Kazakhstan, but a dedicated AI Law and existing provisions of the Criminal Procedure Code, the Data Protection Law, and the Constitution may be construed to regulate or constrain AI use in that context.

At the same time, Kazakhstan has taken steps to engage with international AI governance standards. In January 2026, the Kazakh Foreign Ministry announced a partnership with UNESCO’s local office to apply UNESCO’s Readiness Assessment Methodology (‘RAM’)—a system for evaluating a country’s level of preparedness for the ethical and responsible application of AI. The RAM was specifically created to help member states implement UNESCO’s Recommendation on the Ethics of AI. Kazakhstan has established a ‘National Stakeholder Team,’ which seeks to bring together relevant government ministries, academia, international partners, and the private sector to help guide its implementation of the RAM.

AI regulations

Law of the Republic of Kazakhstan ‘On Artificial Intelligence’ (No. 230-VIII) (17 November 2025)

Kazakhstan’s Law ‘On Artificial Intelligence’ (the ‘AI Law’) defines:

• ‘Artificial intelligence’ as ‘the functional ability to simulate cognitive functions characteristic of humans, yielding results comparable to or exceeding those of human intellectual activity.’
• An ‘artificial intelligence system’ is ‘an information technology system that operates on the basis of one or more artificial intelligence models.’
• An ‘Artificial intelligence model’ is a ‘software product developed to perform specialized tasks and capable of adapting to changing conditions, learning from accumulated experience, and optimizing the processes and results of its activities.’

The AI Law’s overarching purpose is to ensure and promote the development and adoption of AI and to ‘improve [the] quality of human life and increase economic efficiency.’ It seeks to accomplish this goal by establishing a legal framework for regulating AI, ensuring transparent and safe use of AI technology, attracting investment in the country’s AI sector, and providing state support for AI research and innovation. The law outlines seven core regulatory principles, each provided for by its own article:

The law also mandates that developers and operators of AI systems do so ‘with due regard for energy efficiency’ and mitigate ‘negative environmental impacts.’ Art. 8(3). They must also take steps to prevent unauthorised access by third parties to personal data, Art. 10(2), and conduct risk management to prevent potential threats, Art. 11(3).

The law vests strategic policymaking authority in the Kazakh government, Art. 12, and assigns regulatory, supervisory, and implementation functions to a new government body—the newly created Ministry of Artificial Intelligence and Digital Development, Art. 13(1). Other state institutions must participate in AI policy implementation, develop AI models and platforms, and develop data libraries. Art. 13(2). They must provide data access to a National AI Platform—operated by a government-designated entity serving as the central technological infrastructure for the collection, processing, storage, and distribution of data libraries and AI services. Arts. 14, 25.

The law classifies AI systems along three axes: risk level, degree of autonomy, and mode of use:

The law also prohibits the creation or use of AI systems in certain contexts. It bans AI systems that employ subliminal or manipulative techniques, exploit individuals’ vulnerabilities, engage in social scoring (except as permitted by law), unlawfully collect personal data, classify individuals by biometric data for discriminatory purposes, determine emotions without consent, or disseminate prohibited content. Art. 17(3).

Article 16 also sets out a detailed catalogue of user rights and corresponding obligations:

• Users have the right to familiarise themselves with the use agreement of any AI system before use begins, and to the protection of their personal data and confidential information processed by the system.
• The law protects a user’s intellectual property rights to works created using AI, so long as there is ‘a creative human contribution to their creation.’
• Text prompts submitted to AI models that are ‘the result of human intellectual creative activity’ are also eligible for copyright status. The law also entitles users to demand explanations from an AI system owner or operator about system activity that affects their legal rights, in the manner provided by the user agreement and any other applicable law.
• Users can also request information about the data on which an AI system made a decision, subject to limitations imposed by the user agreement and other laws related to the protection of personal data.
• Importantly, users have the right to refuse to interact with an AI system entirely, unless an interaction is compelled by another law.
• On the obligations side, users must use AI systems within the limits of their granted scope of access, and must comply with all applicable rules and safety measures relating to the use of AI systems.

Owners and operators of AI systems have the right to set terms and limits on use of their systems and take measures to protect their rights to their systems (Article 15(1)). However, creative works are only entitled to copyright protection if their creation featured ‘a human creative contribution’ (Article 23). They are also obliged to conduct ongoing risk management, maintain certain documentation, ensure system safety and reliability (including protecting against unauthorised third-party access), and provide user support (Article 15(2)).

To ensure the transparent use of AI, the law mandates that users be informed when goods, works, or services are produced using AI, and that AI outputs be labelled in machine-readable form with accompanying visual or other warnings. Owners and operators of AI systems are responsible for informing users of these requirements. Compensation for damage caused by AI systems is determined by Kazakhstan’s Civil Code. According to Artificial Intelligence and Digital Development Deputy Minister Dmitry Mun, this provision imposes ‘liability for deepfakes.’

Guidelines for practitioners

As at July 2026, there have been no guidelines for practitioners issued in Kazakhstan on the responsible use of AI. In September 2025, the Scientific and Advisory Council of the Kazakh Republican Bar Association (‘RBA’) approved a formal position paper on the use of AI in the practice of law. The position paper was prompted by an inquiry from the Ministry of Justice’s Committee on Registration Services and Organization of Legal Services. After surveying the positions of foreign bar associations—including the American Bar Association, the International Bar Association, the Council of Bars and Law Societies of Europe, the European Bars Federation, and the Law Society of England and Wales—the RBA concluded that the existing legislation, principally the Law ‘On Advocacy and Legal Assistance’, together with the Kazakh Code of Professional Ethics for Lawyers ‘fully cover the duties of a lawyer and the requirements of providing legal assistance using artificial intelligence’ which include: confidentiality, professional competence, professional development, accuracy, quality control, disclosure, independence of professional judgment, and working in the client’s best interest. It found that no additional regulations related to AI and legal practice were necessary at that time.

As at July 2026, Kazakhstan has not formally adopted or otherwise engaged with the UNESCO Guidelines for the Use of AI Systems in Courts and Tribunals (2025).

Criminal procedure rules

Criminal Procedure Code of the Republic of Kazakhstan (2014)

The Criminal Procedure Code of the Republic of Kazakhstan 2014 imposes a series of obligations that may be relevant to the use of AI in criminal proceedings in Kazakhstan:

Digital Code (2026)

Kazakhstan’s Digital Code (signed January 9, 2026, effective on or about July 11, 2026) introduces significant provisions relevant to AI in criminal law once operative. These include a right to appeal decisions made by algorithmic and automated systems with the involvement of authorised specialists; a prohibition on fully automated decisions resulting in discrimination; and a requirement that biometric identification be permitted only where legislation provides for it. Citizens are also guaranteed the right to delete, depersonalise, or suspend the processing of personal data, and the Vice Minister of Artificial Intelligence has stated that service providers must comply with deletion or processing suspension requests within 15 days or face sanctions.

The Code introduces the concept of a ‘fully automated decision’, defined as a decision made without human participation in evaluating the circumstances or confirming the result, where allowed by law or agreement. Individuals subject to such decisions are granted the right to be informed that an algorithmic system has been applied, to receive an explanation of the key factors and criteria that influenced the decision (without disclosure of algorithms, source code, or legally protected information), and to request a human review where the decision has legal consequences or may affect their rights or legitimate interests. The Code also regulates smart contracts, requiring that each smart contract include a dispute-resolution mechanism under which the final decision is made by a human, in accordance with procedures established by Kazakh law or by agreement between the parties. As at July 2026, there is also ongoing work to align existing legislative acts with the Digital Code.

According to Temirlan Ziyatov, chairman of the QazAI Association for the Development of Artificial Intelligence, the Digital Code is the first instrument that ‘systematically regulates data, algorithms, and automated systems’ in Kazakhstan, with requirements for ‘transparency, non-discrimination, and control of decisions’. Ziyatov cautioned, however, that the biggest risk is that rules are adopted ‘on paper’ but are not backed by ‘practical tools, clear guidance, or pilot projects’, and noted a shortage of qualified specialists capable of interdisciplinary work at the intersection of AI and law.

Data protection legislation

Law of the Republic of Kazakhstan, ‘On Personal Data and Its Protection’ (No. 94-V), (adopted 21 May 2013)

Kazakhstan’s Law on Personal Data and Its Protection 2013 establishes the legal framework for the collection, processing, storage, and protection of personal data. Several of its provisions are directly relevant to the deployment of facial recognition tools and other AI systems that process personal data in criminal proceedings.

The law defines:

• ‘Personal data’ as ‘information . . . about a subject of personal data, supplemented by one or more personal data identifiers’. Personal data is broadly categorised into publicly available data and restricted data, with restricted data subject to heightened confidentiality protection.
• ‘Biometric data’ is defined as ‘personal data that characterize the physiological and biological features of the subject’ which ‘may establish [their] identity’. This definition likely encompasses facial recognition data.

As a general rule, the collection and processing of personal data requires the consent of the individual, or their legal representative. For consent to collection of personal data to be valid, it must include: the subject’s name, the time period for which the consent is valid, the list of data to be collected, whether the data will be transferred to third parties, and whether cross-border transfers of the information will occur. Any processing must be limited to ‘specific, predetermined and legitimate objectives.’ Any data that exceeds the scope of consent may not be processed.

Critically, however, the law carves out a broad exception for criminal proceedings. Article 9 allows the collection and processing of personal data without the individual’s consent when carried out ‘by law enforcement agencies, courts and other authorised public authorities that bring and hear cases of administrative offenses, and enforcement proceedings.’ This provides the principal legal basis for the use of AI surveillance tools by police, prosecutors, and courts without obtaining individual consent.

Separately, the law exempts from its scope entirely the collection, processing, and protection of personal data carried out ‘in the course of intelligence, counterintelligence, operational and search activity, as well as implementation of security measures to ensure the safety of protected persons and objects.’ This exclusion could be used to shield certain AI-assisted surveillance activities from the law’s protections altogether.

Cybersecurity laws

As at July 2026, Kazakhstan does not have specific legislation directed at cybersecurity. However, the 2025 Law ‘On Artificial Intelligence’ (discussed above) mandates that developers and operators of AI systems must take steps to prevent unauthorised access by third parties to personal data, and conduct risk management to prevent potential threats.

Human rights

Constitution of the Republic of Kazakhstan (1995)

The Kazakh Constitution outlines several fundamental rights that may limit or regulate the use of AI in criminal proceedings:

Also relevant are the policy-making powers the Constitution grants to Kazakhstan’s president and prime minister. Under Article 45, the president can issue binding decrees and decisions, ‘for the purpose of implementation of the Constitution and laws’. Article 69 provides an almost identical power to the Prime Minister of Kazakhstan, with the caveat that any decrees issued by that office may not contradict enacted legislation, or the decrees and orders of the President.

As to the judiciary, Article 77 governs the administration of justice in Kazakhstan:

1. First, a judge must be independent and ‘subject only to the Constitution and the law’. They are ‘not . . . answerable for individual cases’, and any ‘interference in the administration of justice’ is illegal.
2. Second, judges must adhere to several core principles when applying the law. Every person is presumed innocent until guilt is established by a final court decision, and any doubt as to guilt must be resolved in the accused’s favour. The accused bears no burden to prove innocence, and no one may be compelled to testify against themselves, their spouse, or close relatives as defined by law, with a similar protection extended to priests regarding confessions. No individual may be subjected to repeated criminal or administrative liability for the same offence, and no one’s lawfully determined jurisdiction may be changed without consent. Everyone is entitled to be heard in court. Laws that create or increase liability, impose new obligations, or worsen a person’s situation cannot apply retroactively, though a subsequent law that abolishes or reduces responsibility for an offence must be applied. Evidence obtained illegally carries no legal force, no conviction may rest solely on a defendant’s own confession, and the application of criminal law by analogy is prohibited.

Outlook

Kazakhstan’s regulatory trajectory suggests continued development of AI governance, though the gap between broad-framework legislation and specific criminal procedure rules is expected to persist in the near term.

The Concept for Artificial Intelligence Development for 2024–2029, adopted by the Government in July 2024, is the overarching strategic document for AI infrastructure, talent, research capacity, and governance. It designates priority sectors including state administration and calls for a dedicated legal framework. The establishment of the Ministry of Artificial Intelligence and Digital Development in September 2025 consolidated AI policy, regulation, digital governance, and infrastructure management (including the National AI Platform) under a single institutional authority. President Tokayev declared 2026 the ‘Year of Digitalization and Artificial Intelligence’, and a unified strategy document (‘Digital Qazaqstan’) is under development as at July 2026.

CASES

As at July 2026, there is no publicly available Kazakh criminal case law that adjudicates the legality of AI use by investigators, prosecutors, or judges in criminal proceedings.