Malaysia

Information uploaded as at July 2026

AT A GLANCE

USE

Malaysia is increasingly incorporating AI into different stages of criminal proceedings. Courts in several states have begun using AI-assisted sentencing tools to guide judicial decision-making. These AI tools were developed on the back of Malaysia’s digitisation of its court systems through programmes, including its E-Court system. Law enforcement agencies have additionally expanded the deployment of facial‑recognition systems with embedded AI capabilities.

Law enforcement

Predictive analytics

Though there is no evidenced use of AI tools for predictive analytics by Malaysian law enforcement agencies, senior figures within the Malaysian police have suggested that there is an interest in the use of AI for predictive policing. For example, in June 2025, the Inspector General of the Royal Malaysian Police identified as a department priority the use of AI to address issues and threats. A few months later, the Deputy Inspector General added that the force intended to use advanced technology to ‘counter manipulation trends by global crime syndicates’ and employ AI to reconstruct crime scenes.

Data review and analysis

In 2018, the Auxiliary Force, a unit of the Royal Malaysian Police, announced a partnership with Chinese AI company YITU Technology to incorporate facial recognition software into body-worn cameras. The CEO of the Auxiliary Force explained in a press release that YITU’s technology was a step toward ‘instant alerts’ notifying police of individuals on criminal watch lists. In 2018, YITU claimed it could identify a person from its database of 1.8 billion humans within three seconds to an accuracy level of 95%.

In January 2019, the Penang Island City Council (Majlis Bandaraya Pulau Pinang or ‘MBPP’) and another local agency upgraded over 1000 CCTV cameras with facial recognition capabilities across the state. Penang’s Chief Minister, Chow Kon Yeow, explained that the CCTV cameras were ‘fully linked with the police operations room’ and issued a warning to wrongdoers that they were ‘being watched’ by the police. The technology, provided by IBM, would allow the police and MBPP to collate data and use analytics to identify offences such as illegal parking and unauthorised use of emergency vehicle lanes.

The Minister for the Federal Territories in 2020 shared his intention to install 5000 CCTV cameras in convenience stores across Kuala Lumpur with the help of an undisclosed private company. Then-police chief Mazlan Lazim said that the cameras would store the faces of alleged criminals and allow his force to trace their identities with the aid of facial recognition technology.

State and local governments continue to invest in CCTV cameras with AI capabilities. In February 2026, the Mukah police in Sarawak installed such cameras on the Pan Borneo Highway.

Prosecutors

As at July 2026, there are no reported instances of prosecutors in Malaysia making use of AI.

Courts

Malaysia has modernised its case management infrastructure, primarily by adopting electronic case management systems with automation elements. After an initial failed attempt in 2002, Malaysian courts began implementing an E-Court system in 2009 that included four modules: Electronic Filing System, Queue Management System, Case Management System, and Court Recording and Transcribing and Video Conferencing System:

This automation of case management processes has modernised Malaysian courts over the past decade. However, on their own, these four systems do not use AI. Rather, they have provided the data foundation for the Malaysian judiciary to train AI models for more sophisticated purposes, such as decision-making support (see below).

Decision-making support

The AI Sentencing System was developed through a collaboration between the Malaysian judiciary and Sarawak Information Systems and integrated into the E-Court system of Sabah and Sarawak, sometimes referred to as the Artificial Intelligence in Court System. The system was built on details in the E-Court Case Management System (see above). In February 2020, the Magistrate Court in Kota Kinabalu used the AI Sentencing System for the first time in cases under the Dangerous Drugs Act 1952. Statements by officials on the AI Sentencing System emphasise that the tool is intended only as a tool to guide legal decision-makers, and should not replace human judgment. For example, in a submission to the UN Office of the High Commissioner for Human Rights, Malaysia described the system as a ‘non-binding guideline for judges’.

In July 2021, the Office of the Chief Registrar of the Federal Court of Malaysia announced the phased implementation of AI-Assisted Sentencing Guidelines in the Sessions Courts and Magistrates’ Courts across Peninsular Malaysia. The system was intended to function solely as a decision-support tool rather than to replace judicial discretion. Judges and magistrates enter relevant case information—including the offender’s age, nationality, occupation, marital status, prior criminal record, and, in drug-related offences, the quantity of drugs involved—after which the system generates a recommended sentencing range. These recommendations are derived from an algorithm that analyses historical sentencing patterns from comparable cases decided by the Sabah and Sarawak courts, where the AI system had first been implemented, to promote greater consistency in sentencing while leaving the final decision to the presiding judicial officer.

Defence

As at July 2026, there are no reported examples of defence counsel in Malaysia making use of AI.

Victims

Victims do not have legal standing in Malaysian criminal proceedings. Article 145(3) of the Malaysian Constitution gives the Attorney General the ‘power, exercisable at his discretion, to institute, conduct or discontinue any proceedings for an offence, other than proceedings before a Syariah court, a native court or a court-martial.’ Further, Article 376(1) of Malaysia’s Criminal Procedure Code states that ‘[t]he Attorney General shall be the Public Prosecutor and shall have the control and direction of all criminal prosecutions and proceedings under this Code.’ Malaysian courts have consistently interpreted these two provisions to mean that the Attorney General has the ‘absolute discretion […] to institute criminal proceedings.’ Foo Tseh Wan v. Toyota Tsusho (M) Sdn Bhd & Anor, [2018] 9 MLJ 664, 681-683.

As at July 2026, however, there are no reported examples of victims in Malaysia using AI tools in criminal proceedings.

TRAINING

As at July 2026, there is no mandatory or systemic AI-related training for law enforcement, prosecutors, courts, and defence counsel in Malaysia. There are, however, ad hoc initiatives to offer training to criminal justice practitioners.

It has been reported that the Department of Syariah Judiciary Malaysia (‘JKSM’) has developed a training module addressing the application and ethical use of AI in the Syariah judicial and legal fields. As part of continuous development efforts, this AI course will be incorporated annually into the JKSM’s Training Operational Plan. The course will also be made available to officers and support staff serving in Syariah Courts throughout Malaysia. However, publicly available information on the module’s content, scope, or implementation remains limited.

The Malaysian Bar has also integrated AI-related content into its Continuing Professional Development programme through accredited Continuing Professional Development live webinars, including sessions such as ‘Artificial Intelligence and the Lawyer — AI and You; AI and Me’ (27 January 2026) and state bar webinars such as ‘Use of Generative AI for Lawyers’ (9 February 2026).

While the Malaysian Ministry of Digital has implemented general AI training programmes for civil servants, there is no evidence that the Ministry or judiciary have created specific training on deepfakes.

REGULATION

As at July 2026, Malaysia is in the process of developing a formal regulatory framework for AI—the forthcoming Artificial Intelligence Governance Bill—which aims to promote responsible and ethical use of AI across sectors. Although no AI‑specific rules govern criminal proceedings, the government has issued general AI governance and ethics guidelines, and several existing legal frameworks may indirectly influence how AI tools are deployed in the justice system. Furthermore, judges and lawyers remain subject to established ethical standards that, while technology-neutral, are relevant to the responsible use of AI in criminal proceedings.

AI regulations

As at July 2026, an Artificial Intelligence Governing Bill is forthcoming. The Bill aims to promote the responsible use of AI, while aligning with international best practices and complementing existing legislation to create a coherent governance framework across sectors. The Ministry of Digital (the Malaysian government ministry responsible for digital policy) is developing the legislation through the National Artificial Intelligence Office (‘NAIO’) using a risk-based approach that covers AI-related harm, incident reporting and ethical principles. On July 10, 2026, the NAIO released a Public Consultation Paper to solicit stakeholder input on the Bill’s proposals. The Ministry of Digital expects to introduce the Bill in the second half of 2026.

Guidelines for practitioners

As at July 2026, guidelines have been issued for lawyers and judges in Malaysia. These guidelines stem primarily from pre-existing professional and ethical codes rather than AI-specific instruments, although lawyers have also received dedicated guidance on generative AI.

Malaysia has also expressed a commitment to global and regional standards on the use of AI, for example, by subscribing to the UNESCO Recommendations on the Ethics of Artificial Intelligence (2023), but as at July 2026, Malaysia has not formally adopted the UNESCO Guidelines for the Use of AI Systems in Courts and Tribunals (2025).

Guidelines applicable to Malaysian lawyers

The Legal Profession Act 1976 and the Legal Profession (Practice and Etiquette) Rules 1978 together establish the ethical framework governing advocates and solicitors in Malaysia. While neither instrument expressly addresses the use of AI, they impose clear duties of honesty, accuracy, candour and good faith in dealings with the court. The Act adopts a broad concept of professional misconduct (e.g., Section 94 provides that ‘misconduct’ includes dishonest or fraudulent conduct in the discharge of duties) which may be engaged where AI-generated or AI-assisted materials are relied on without proper verification. The Rules further prohibit misleading the court, misquoting authorities or evidence, referring to unproven facts, or engaging in deception, and require advocates and solicitors to conduct proceedings with fairness and candour. Taken together, these obligations mean that lawyers must exercise particular care to ensure the accuracy, completeness and integrity of any submissions or materials generated or assisted by AI in criminal proceedings. Similar duties apply to lawyers in Sabah and Sarawak under separate but analogous professional regimes. For instance, the legal profession in Sabah follows ethical guidance from the Advocates Ordinance and the Advocates Rules of 1988, which have similar duties to ‘candour’ and ‘fairness.’ Sarawak has adopted Advocates Rules that outline duties to ‘candour’ and ‘fairness.’

On 3 July 2025 the Malaysian Bar Association published Circular No. 242/2025 on the responsible use of generative AI. The Circular makes clear that AI may be used by advocates and solicitors only as a limited productivity aid, never as a source of legal authority or judgment. Practically, lawyers may use AI tools for low‑risk tasks such as brainstorming, early drafting, formatting, simplifying text, or preparing internal training materials, but must not use it to give legal advice, verify or cite case law or legislation, conduct jurisdiction‑specific legal analysis, draft pleadings or contracts without full legal review, or predict case outcomes. The circular stresses that no confidential, privileged, or personal data may be entered into AI tools, and that all outputs must be treated as unverified and independently checked against authoritative sources.

Guidelines applicable to Malaysian judges

Malaysia’s Judges’ Code of Ethics 2009 (‘Judges’ Code’) is the principal ethical instrument governing judges of Malaysia’s superior courts, namely the Federal Court, the Court of Appeal, and the High Court. The Judges’ Code establishes a framework that, while not AI‑specific, requires judges to exercise independent judgment, act with diligence, avoid bias, and uphold the integrity of the judicial process. These obligations have several possible implications for the use of AI tools in criminal proceedings. For example, the duty of independence may be understood as requiring judges to form their own assessment of the facts and the law, rather than fully deferring to AI recommendations. Similarly, the duty of diligence could be read as calling for caution where AI‑generated outputs are used, including an expectation that such outputs are checked and verified before any reliance is placed on them. Comparable standards apply in subordinate courts under the Judicial Officers’ Code of Ethics 2019, which closely mirrors the Judges’ Code and imposes analogous duties of independence, diligence, integrity and fairness on Sessions Court Judges, Magistrates and Registrars.

Criminal procedure rules

Malaysia does not have criminal procedure rules specifically governing the use of AI, including deepfake evidence, in criminal procedure. In the absence of AI‑specific rules, the use of AI in criminal proceedings may be assessed in light of existing laws on electronic, computer‑based, and digital evidence. These laws were not designed with AI systems in mind but are capable of being interpreted and applied to AI-enabled tools and outputs in the criminal procedure context.

Evidence Act, 1950

The Evidence Act 1950 includes provisions that allow evidence produced by a computer to be admitted in legal proceedings. Section 3 defines a computer, in relevant part, as ‘an electronic […] or other data processing device, […] performing logical, arithmetic, storage and display functions, and includes any data storage facility or communications facility directly related to or operating in conjunction with such device or group of such interconnected or related devices.’ Section 62 states that a ‘document produced by a computer is primary evidence’ and can be produced to be inspected by the court. Sections 90A, 90B, and 90C provide further details about the admissibility and weight of computer evidence. Although these provisions predate modern AI systems and do not expressly refer to AI technologies, AI-generated or AI-assisted outputs could potentially be treated as computer‑produced documents for evidentiary purposes.

Criminal Procedure Code, as at September 2017

The Criminal Procedure Code similarly does not address the use of AI, but contains broad provisions related to computer evidence that can be accessed by an Inspector conducting a criminal investigation. Section 116B states that ‘[a] police officer not below the rank of Inspector conducting a search under this Code shall be given access to computerised data whether stored in a computer or otherwise’ and that the results of this computerised investigation ‘shall be admissible in evidence notwithstanding any other provisions in any written law to the contrary.’ While the provision does not distinguish between different forms of computerised analysis, it is sufficiently broad to potentially encompass AI-assisted data processing used during investigations.

Computer Crimes Act 1997; Communications and Multimedia Act 1998; and Cyber Security Act 2024

While the Evidence Act and Criminal Procedure Code govern the legal procedures that surround the use of computerised evidence, the Computer Crimes Act 1997, Communications and Multimedia Act 1998, and Cyber Security Act 2024 operate at a more general level by regulating conduct involving computers, networks, and electronic systems. These statutes do not specifically relate to the use of AI in criminal procedure or other legal proceedings. Rather, they address offences committed using computer, cyber, and electronic means, which may include or be facilitated by AI technologies. As with the Evidence Act and Criminal Procedure Code, these Acts are technologically broad and do not expressly regulate AI systems or their use by law enforcement or the courts but may become relevant where AI tools are implicated in cyber‑related offences or security risks.

Data protection legislation

Personal Data Protection Act, 2010 and Personal Data Protection (Amendment) Act, 2024

The Personal Data Protection Act 2010, as amended by the Personal Data Protection (Amendment) Act 2024, is Malaysia’s primary data protection framework. The Act applies to the processing of personal data in the context of commercial transactions. While this means the Act may be relevant to certain AI activities with a commercial character (such as the development or training of proprietary AI systems for commercial purposes and later deployed in criminal justice settings), its application to criminal proceedings may be limited in light of the Act’s exemptions for data processing relating to prosecutions and court orders or judgments.

Cybersecurity laws

Cyber Security Act, 2024

As introduced above, the Cyber Security Act 2024 (‘CSA’) is Malaysia’s principal cybersecurity law and establishes a regulatory framework focused on the security and resilience of critical information systems. The Act applies to entities designated as National Critical Information Infrastructure (‘NCII’) and to providers of certain prescribed cybersecurity services. In this respect, the CSA is most likely to be relevant to courts and prosecution services in their capacity as operators of government IT systems, where the information systems they operate are designated as critical infrastructure. In such cases, NCII-designated courts or prosecution services would be subject to operational cybersecurity obligations (such as risk assessments and incident notification), which may influence how AI tools are implemented and secured within institutional IT systems. However, the CSA does not directly regulate the use of AI tools in criminal proceedings as such.

Human rights

Federal Constitution of Malaysia

The Malaysian Federal Constitution 1957 is the supreme law of Malaysia. As such, all practices relating to criminal proceedings, including the development and use of AI tools, must comply with its guarantees. Several constitutional provisions may be relevant in this context:

• Article 5 (liberty of the person): provides that ‘[n]o person shall be deprived of [their] life or personal liberty save in accordance with law.’ This provision may be engaged where AI tools influence decisions relating to arrest, detention or imprisonment. Furthermore, although the Constitution does not contain an express, freestanding right to privacy, Malaysian courts have recognised that the right to personal liberty under Article 5 encompasses aspects of privacy, which may be relevant where AI systems process personal or biometric data.
• Article 8 (equality): provides that ‘[a]ll persons are equal before the law and entitled to the equal protection of the law’. This provision may be relevant where algorithmic tools used in criminal justice processes risk producing biased or discriminatory outcomes.
• Article 13 (rights to property): provides that ‘[n]o person shall be deprived of property save in accordance with law’. This provision may be relevant where enforcement measures and the use of AI tools relate to the seizure or forfeiture of assets.

Malaysia is not a signatory to the International Covenant on Civil and Political Rights 1966. Malaysia, however, is a signatory to the UN Convention on the Rights of the Child (CRC). The CRC includes provisions related to privacy that are relevant to the development and use of AI. For example, Article 16(1) provides that ‘[n]o child shall be subjected to arbitrary or unlawful interference with his or her privacy, family, home or correspondence, nor to unlawful attacks on his or her honour and reputation.’ The use of personal data to train AI systems, for example, with passive collection through social media, creates risks to a child’s privacy that encroach on this provision of the CRC.

Outlook

Malaysia is focused on integrating AI into its governance frameworks. Malaysia’s National Artificial Intelligence Office (‘NAIO’) possesses the express goal of ‘transforming Malaysia from an AI consumer into an AI producer by building a robust AI ecosystem that will accelerate economic growth, improve public services, and ensure inclusive and equitable access to AI where no one is left behind’. Part of NAIO’s task has been to develop the national AI regulatory framework, including the Artificial Intelligence Governance Bill referenced above. The NAIO is currently soliciting stakeholder input on the Bill’s proposals, and the Ministry is expected to introduce the Bill in the second half of 2026. Current publications regarding the framework do not identify any provisions related to the use of AI in criminal procedure. Thus, it is not clear if there are existing legislative priorities related to integrating AI into Malaysia’s criminal legal system.

Alongside the development of the Bill, the Government is updating key technology related legislation as part of a broader effort to ensure a safe AI ecosystem. Digital Minister Gobind Singh has indicated that this includes amendments to the Personal Data Protection Act 2010, implementation of the Cyber Security Act 2024 and the Online Safety Act 2025, updates to the Evidence Act 1950 and Contracts Act 1950, and the introduction of a Cybercrime Bill to replace the Computer Crimes Act 1997.

CASES

As at July 2026, there have been a number of cases before Malaysian courts where judges have used AI, and disclosed such use.

Disclosed uses of AI by judges

One prominent case on the use of AI in Malaysia’s legal system is the Denis v Public Prosecutor [2020] 2 SMC 381 (Magistrate Court, Kota Kinabalu). In Denis, the defendant was convicted of possessing 0.01 grams of methamphetamine under the Dangerous Drugs Act of 1952. Malaysia’s AI sentencing system suggested a 10-month prison sentence, and the court increased this sentence to 12 months to ‘not only give a clear message to the offenders out there but the would-be offender as well.’ However, on appeal, the High Court ultimately reduced this sentence without explanation.

In Public Prosecutor v. Adrian Kheung Peng Yin [2021] MLJU 2994 (Magistrate Court, Kuala Lumpur), a defendant was convicted of violating Section 323 of the Malaysian Penal Code (‘voluntarily causing hurt’) and the AI sentencing system proposed a fine within the range of RM 800.00 to RM 1000.00. The Court, however, did not follow the AI system and increased the fine to RM 1800.00.

Conversely, some decisions have shown a willingness to support AI sentencing proposals. In Sukerdi bin Damit v Polis Diraja Malaysia (PDRM) Sandakan [2025] MLJU 3303 (High Court, Sandakan), the appellant was convicted of possessing 0.07 grams of methamphetamine and sentenced to serve 10 months in prison. One ground of appeal was that the AI-recommended 10-month sentence was excessive. The appellate court upheld the sentencing decision, stating that ‘the learned Magistrate did refer to and correctly applied the relevant principles of sentencing before coming to his conclusion on the sentencing. He had taken into account all relevant considerations. There is no reason for this Court to disturb the sentence’.