Türkiye
Tool
Tasks
User
Scope
Training
Regulation
Insight
AT A GLANCE
Türkiye is actively exploring AI in its justice system under its National AI Strategy (2021–2025) as updated in the National AI Action Plan (2024-5). Early-stage AI is used in law enforcement (KGYS for criminal surveillance, Trafidar for traffic surveillance, ASENA to identify suspicious transactions) and predictive analytics (reoffending risk assessment). Within prosecution and the courts, AI supports case classification, indictment inconsistency detection, workload distribution, and legal text summarisation although no tools are as yet fully deployed in courtrooms. Lawyers use generative AI for drafting on an individual basis. Training initiatives are ongoing via the Justice Academy and regional bar associations.
Türkiye does not have legislation targeting the use of AI in courts, which is regulated by general criminal law and procedure, data protection legislation and cybersecurity laws and by AI-specific recommendations and guidance issued by Türkiye’s data protection authority and regional bar associations. An AI Bill is before parliament but as at September 2025, had not been enacted.
USE
As at August 2025, AI implementation within criminal justice in Türkiye remains in development rather than deployment. Türkiye does not have a well-established AI tool that is actively and officially used in criminal proceedings nationwide. However, Türkiye has begun to use early-stage AI tools to assist law enforcement and the judiciary.
Law enforcement
Data review and analysis
Türkiye uses a two-step infrastructure for law enforcement: (1) surveillance and data collection; followed by (2) AI-powered analytical systems.
|
Surveillance and data collection |
The Urban Safety Management System (KGYS) is Türkiye’s main infrastructure for city-wide surveillance. KGYS integrates video surveillance, automated license plate recognition, and data links across municipalities. Data generated by these systems has been confirmed to be used in criminal investigations, including the identification of protesters during mass demonstrations and in terrorism-related cases. Türkiye’s Minister of Internal Affairs announced in local media that by the end of 2025, all police officers would be equipped with body cameras featuring facial recognition technology. |
|
Trafidar (produced by Radarsan) is a national surveillance system used by the Turkish government to automate traffic enforcement, speed and license plate monitoring. It is an automated speed enforcement system. The Modular Traffic Radar System, developed using AI, machine learning and big data technologies, has a wide range of applications in autonomous and smart systems. Within a period of five days in June 2025, fines totalling 526,645 Turkish liras were issued to 224 vehicles that exceeded the speed limit. |
|
|
AI-powered analytical systems |
The nationwide CBS Organizational Prediction Project, discussed below, uses AI to identify potential links between new case entries and previously classified terrorist organizations. |
|
The Analysis System Narcotics Network (ASENA) launched in 2021, is a nationwide AI-powered platform used by police to combat narcotics crime by identifying suspicious transactions. It has reportedly supported thousands of criminal investigations. ASENA uses machine learning and pattern-recognition algorithms to analyse vast datasets (including phone records, financial transactions, and geolocation data), detect anomalies, and identify hidden connections between individuals, routes, and events. The system also conducts network analysis to map suspected criminal organisations. ASENA ‘learns’ autonomously based on arrest and investigation data, processing a vast number of queries (around 300 million) with an accuracy rate of approximately 92%. It autonomously generates insights and classifications that can directly trigger criminal investigations. All actions and queries within the system are recorded and monitored. In 2022, it was reported that the system was used by 3,100 police officers, 650 of whom underwent training to work with ASENA. As at August 2025, the ASENA system continues to be in use, and in March 2025 Turkish authorities reported the use of the system in a narcotics operation in Şırnak, where the system contributed to the detection and seizure of 86.6 kilograms of drugs. |
Predictive analytics
In 2025, authorities introduced a recidivism risk assessment system, designed to forecast the likelihood of repeat offending, particularly in cases of domestic violence. The system assists in determining preventative measures and escalation risks, and its predictive accuracy (of around 75%) reportedly exceeds that of comparable European tools, such as the UK’s HART.
Prosecutors
Case management
The CBS Organizational Prediction Tool, an initiative put forward by the Turkish Ministry of Justice and integrated into the National Judiciary Informatics System (UYAP), is a system designed to automatically identify potential associations between new case entries and previously classified terrorist organizations in the national judiciary database. By algorithmically linking individuals or cases to terrorist organizations, the system may effectively “tag” persons even before a judge has reviewed the case.
Human errors and deficiencies during the entry of organizational information into public prosecutor’s office case files, along with incomplete data entries, had been causing significant problems. The artificial intelligence model we developed has provided a solution to this issue. Our project automatically matches newly entered data with the information on terrorist organizations already registered in the system, thereby minimizing data entry errors.
Artificial intelligence projects are not intended to eliminate the judge's discretion. The authority to make decisions will still rest with the judge. These projects will reduce the workload of judicial personnel and enhance objectivity and efficiency in decision-making processes.
There are reported incidents of Turkish lawyers making unofficial use of commercial generative AI tools such as ChatGPT for drafting purposes.
Another important issue is that our colleagues who have spent many years in the legal profession must also adapt to this change. Therefore, as the Istanbul, Ankara, and Izmir Bar Associations, we will take steps to accelerate this process through in-service trainings.
REGULATION
Türkiye does not have legislation targeting the use of AI in court proceedings. It is regulated by existing laws, including criminal procedure, data protection and cybersecurity legislation. Türkiye’s data protection authority, the KVKK, has issued recommendations on the regulation of AI and the Ankara Bar has issued non-binding guidance to its members. Türkiye’s regional bar associations have been early advocates of AI regulation. In 2019, the Istanbul, Ankara and Izmir Bars published ‘Law in the Age of AI’, Türkiye’s first comprehensive legal analysis of AI, which preceded the launch of Türkiye’s national AI strategy four years later. General draft AI legislation was introduced in 2024 but as at August 2025, has not yet been enacted.
Guidelines for practitioners
Attorneyship Law (Law No. 1136) and Code of Professional Conduct
Attorneys must at all times adhere to their obligations under the Attorneyship Law and the Union of Turkish Bar Associations’ Code of Professional Conduct Code of Professional Conduct. These instruments impose strict duties of confidentiality, independence, and accountability. Even though they do not refer to AI, they may apply to its use in court and may require that any lawyer using AI must ensure client data remains confidential and outputs are verified before use.
The Ankara Bar’s Guide on the Use of AI in Legal Practice (2024)
The Ankara Bar’s Guide on the Use of AI in Legal Practice provides guidance on the use of AI tools. The Guide discusses risks such as algorithmic bias resulting in discriminatory or unfair outcomes, AI hallucinations creating outputs that are nonsensical or altogether inaccurate, deepfakes(i.e. AI-manipulated videos or audio recordings of a person's likeness so that they appear to do/say something they never did), and black box systems whose internal workings are not transparent. The Guide emphasises that legal practitioners should be made aware of these risks. The Guide itself is not binding, but it advises lawyers to ensure client confidentiality when using AI tools that may store input data and reuse it. Lawyers should verify AI outputs to eliminate the risk of providing misleading information, and they should maintain personal accountability in line with their binding professional obligations under the Attorneyship Law (see above).
Law on Judges and Prosecutors (Law No. 2802)
Judges and prosecutors are bound by the Law on Judges and Prosecutors, which establishes judicial independence and provides tenure protection, i.e. judicial and prosecutorial security, meaning that they cannot be removed from their position, discharged, pensioned off, or transferred to another location without their consent. Moreover, the Law emphasises that judges must render their judgements based on their conscientious conviction. Even though AI is not mentioned, this may require that they not delegate their decision-making authority to AI tools or rely on them without independent assessment.
Criminal procedure rules
Turkish Criminal Procedure Code (Law No. 5271)
The Turkish Criminal Procedure Code contains several procedural safeguards that indirectly regulate the use of AI even though the Code does not refer to AI:
- Admissibility of evidence (articles 206–217): Evidence must be lawfully obtained, reliable, and relevant. AI-generated or AI-analysed evidence (e.g., facial recognition, predictive analytics) must meet these standards. Judges have discretion to exclude evidence if its origin or reliability is questionable, which might be important for black-box AI systems.
- Procedure for testimony and interrogation (articles 147–148): Defendants must be able to understand and challenge the evidence against them. If AI systems are used, their logic and outputs must be explainable and accessible to defence counsel.
- Protection of personal data in proceedings (articles 80, 209): AI systems must comply with data protection rules, especially when handling sensitive data like biometric identifiers or criminal records.
Article 134 of the Criminal Procedure Code permits the search and provisional seizure of computers. A reform proposal for this article is articulated in the Ankara, Istanbul and Izmir Bars’ joint 2019 report on ‘Law in the Age of AI’ (see above) to reflect the role AI tools may play in evidence collection.
Turkish Criminal Code (Law No. 5237)
The Turkish Criminal Code sets out the substantive criminal law in Türkiye and includes provisions that indirectly affect AI use in criminal justice:
- Unlawful data processing (articles 135–140): These articles criminalize unauthorized collection, recording, and dissemination of personal data. If an AI system used in criminal investigations processes personal data without legal basis or consent, it may violate these provisions.
- Forgery and manipulation (articles 204–212): AI-generated evidence or manipulated digital content (e.g., deepfakes) could fall under forgery or digital fraud if used to mislead judicial authorities.
Law on the Regulation of Broadcasts via Internet and Prevention of Crimes Committed through Such Broadcasts (No. 5651) (Internet Law)
The Internet Law is one of the key laws that indirectly governs offences committed through AI use, as it regulates internet-based content and activities that may involve AI systems. The Internet Law addresses four main areas:
- setting out the civil, criminal, and administrative responsibilities of internet actors, including content, hosting, access public use, and social network providers;
- establishing procedures for restricting access in relation to certain crimes, with particular emphasis on emergency measures;
- governing notice-and-takedown processes, enabling the removal or blocking of unlawful online content, especially in cases involving violations of personal rights; and
- regulating filtering methods and the broader monitoring of online activities.
Data protection legislation
Law on Protection of Personal Data (No. 6698)
Türkiye’s Law on the Protection of Personal Data regulates all processing of personal data and would therefore cover AI systems handling personal data in criminal investigations or trials. For instance, if law enforcement uses facial recognition to identify suspects, that processing must have a legal basis under the Law on Protection of Personal Data (such as a statutory authorization or consent) and must comply with principles of lawfulness, accuracy, purpose limitation, data minimization, and security.
KVKK Recommendations
The Turkish Personal Data Protection Authority (KVKK) has issued non-binding but influential Recommendations on the Protection of Personal Data in the Field of Artificial Intelligence addressed to ‘developers, manufacturers, service providers and decision makers in the field’ of AI, which will also regulate the use of AI by the police and in the courts. These emphasize several principles, including that AI systems must respect fundamental rights and avoid discriminatory outcomes. Only necessary personal data should be processed, and its quality and ‘the accuracy of the developed [AI] model’ must be continuously monitored. AI systems should be explainable, and their decision-making processes should be auditable. Developers must assess and mitigate risks such as bias, discrimination, and privacy violations. The design of AI applications should be ‘ethically and socially oriented’, ideally involving academic and civil society input. The Recommendations promote measures like anonymization of personal data while processing it, adhering to the principle of proportionality, and giving individuals the right to opt-out of automated processing.
Consumer Protection Law (No. 6502) and E-Commerce Law (No. 6563)
The Consumer Protection Law and the E-Commerce Law govern digital transactions and consumer rights, including legal services. Legal platforms have an obligation to ensure that digital agreements are legally binding and verifiable, meeting standards under the Turkish Commercial Code and E-Commerce Law. Platforms must align with the Law on Protection of Personal Data (see above) for personal data processing, including consent management, breach notification, and cross-border data transfers.
Cybersecurity laws
Law on Cybersecurity (No. 7545)
Enacted in March 2025, the Law on Cybersecurity is Türkiye’s first comprehensive framework for digital security. It is not directed specifically at lawyers, but imposes obligations that directly affect law firms and legal tech providers operating in cyberspace and may therefore also apply to the use of AI tools in court. The obligations include:
|
Certification and licensing |
Cybersecurity products and services used by law firms must be certified by the Cybersecurity Directorate. Unauthorised use or sale of uncertified tools may result in penalties. |
|
Data protection and confidentiality |
Legal professionals handling sensitive client data must ensure compliance with strict confidentiality rules. Breaches or unauthorised disclosures can lead to imprisonment (up to five years). |
|
Critical infrastructure definition |
Cybersecurity products and services used by law firms must be certified by the Cybersecurity Directorate. Unauthorised use or sale of uncertified tools may result in penalties. |
|
Oversight and cooperation |
Legal service providers must cooperate with the Cybersecurity Directorate, including sharing logs and incident data when requested. |
Human rights
The Turkish Constitution protects core rights relevant to AI use, including the right to a fair trial (article 36), the principle of legality in criminal law and the presumption of innocence (article 38), and the right to protection of personal data (article 20).
Regional instruments provide additional guidance. Türkiye is a party to the Council of Europe’s European Convention on Human Rights, which guarantees the right to a fair trial (article 6), to privacy (article 8), and non-discrimination (article 14). These rights apply directly to judicial proceedings and may extend to the use of AI in that context.
In addition, the European Ethical Charter on the use of AI in the judicial systems and their environment, adopted by the Council of Europe’s European Commission for the Efficiency of Justice (CEPEJ) in 2018, sets out five general principles on the use of AI. First, the design and implementation of AI tools and services must be ‘compatible with fundamental rights’. Second, the ‘development or intensification of any discrimination between individuals or groups of individuals’ must be prevented. Third, ‘certified sources and intangible data with models conceived in a multi-disciplinary manner, in a secure technological environment’ must be used for the ‘processing of judicial decisions and data’. Fourth, ‘data processing methods’ must be ‘accessible and understandable’, and external audits must be authorised. Finally, users must be ‘informed actors and in control of their choices’.
In 2024, the Center for AI and Digital Policy called for Türkiyy to sign the Council of Europe Framework Convention on Artificial Intelligence, Human Rights, Democracy and the Rule of Law but Türkiye had not announced its position as at August 2025. The treaty requires states to ensure that AI systems are not used to undermine ‘respect for judicial independence and access to justice’.
The protections of the right to a fair trial and the right to privacy in articles 14 and 17 of the International Covenant on Civil and Political Rights and articles 16 and 40 of the Convention on the Rights of the Child, both ratified by Türkiye, may also be relevant.
Outlook
Government AI strategy
Türkiye’s government strategy and development plans emphasise the development, use and regulation of AI, including in the justice sector. Its National Artificial Intelligence Strategy 2021-2025, adopted in 2021, was the country’s ‘first national strategy in this field’. The strategy’s aims included boosting training, employment, and research in the field of AI, and ‘increasing the maturity of [the] country’s AI ecosystem’. The strategy also called for the establishment of an ‘appropriate ethical and legal framework for AI’ aligned with international standards.
Türkiye’s Twelfth Development Plan (2024-2028), approved in 2023, specified that ‘the use of artificial intelligence in justice services’ would ‘be enhanced’, ‘[d]ecision support systems’ would be ‘strengthened by taking into consideration the opportunities’ brought by AI, and that the ‘ethical and legal framework’ of AI applications would be constituted. Türkiye’s 2024-2025 National Artificial Intelligence Action Plan, issued in 2024 to update the National AI Strategy in light of the Twelfth Development Plan, named the Ministry of Justice and the Union of Turkish Bar Associations as the stakeholders responsible for preparing a legal assessment guide for AI applications. The Union of Turkish Bar Associations had been active in AI-related legal work for over a decade and established a dedicated AI and Law R&D Unit in 2019.
The Ministry of Justice’s Judicial Reform Strategy 2025-2029, launched in January 2025, confirmed that an ‘institutional framework’ would be established to develop AI applications and to ‘define their ethical and legal boundaries’. The Ministry referred to plans for AI-based ‘recommendation systems’ to be ‘developed to support judicial proceedings’, for AI to be ‘utilised for real-time document analysis’, form-filling, verification of expert reports, tracking of correspondence, and ‘inter-agency information sharing’. In the context of criminal proceedings, the strategy proposed the introduction of AI ‘algorithms to assess applications for the deletion or correction of criminal records’.
Draft AI legislation
In 2024, an AI Bill for the regulation of AI in Türkiye was submitted to Türkiye's parliament, the Grand National Assembly of Türkiye. The AI Bill seeks to regulate AI based on principles of safety, transparency, equality, accountability, and privacy. It introduces the concept of ‘AI operators’ (including providers, users, importers, and distributors) and requires them to comply with these principles. In terms of enforcement, the Bill provides that the ‘relevant supervisory authorities’ should monitor compliance and detect violations for which AI operators may be fined a percentage of their annual turnover. The Bill defines AI as computer-based systems capable of human-like functions such as learning, reasoning, and perception. It requires risk assessments and conformity checks for high-risk AI systems. These must be registered with the relevant supervisory authorities and pass a conformity assessment prior to deployment.
As at September 2025, the Bill is under parliamentary committee review. If approved, it will be debated and voted on by the Grand National Assembly. Separately, the Grand National Assembly has established an AI Research Commission to study best practices and help develop the regulatory framework for AI, including in criminal justice. The Commission consulted public institutions, universities, and actors in the private sector. Stakeholders involved in informing the debate included the Better Justice Association, a Turkish thinktank. Burhan Uyan, one of its principal members, has emphasized the need for trustworthy AI tools and the dangers of relying on untested models.
We need to make sure that a human with the relevant expert experience remains in charge of the process at the final stage.
