At the end of April this year, it was reported that around 10,000 enhanced closed-circuit television (CCTV) cameras had been installed across the city of Kuala Lumpur, each intended to support law enforcement in tackling crime rates and ensuring public safety.
The artificial intelligence (AI)-powered camera network, in development since 2020 and costing RM500mil to operate, has already shown results, with Inspector-General of Police Tan Sri Mohd Khalid Ismail saying in a June report that it has assisted police in solving cases and carrying out their duties.
Such cameras are strategically placed throughout the city, specifically targeting high-activity areas, including intersections, major roads, commercial centres, and public areas.
According to Kuala Lumpur police chief Comm Datuk Fadil Marsus in the April report, the system has increased efficiency in tackling crime by about 50%.
He adds in a separate April report that the system’s implementation has led to a decrease in street crime in 2025, resulting in a 57.6% decrease in snatch thefts – from 368 in 2024 down to 156 in 2025.
Adding AI to CCTV cameras reportedly allows them to detect actions in real time and the ability to function as a geofence, creating a virtual perimeter around a specific area in the city. This allows authorities to monitor activity within the zone, track movements, and detect persons or vehicles of interest in real time.
The system also comes with facial recognition features, the ability to analyse behaviour and identify suspicious movement, and can detect early signs of incidents, such as floods, accidents, emergencies and traffic disruptions.
All of this is then monitored at the Kuala Lumpur Command and Control Centre (KLCCC), which will be automatically alerted by the system if any potential incidents are detected.
From there, the KLCCC will coordinate responses between various agencies, such as emergency services like the Malaysian Fire and Rescue Department (Bomba) and the Royal Malaysia Police (PDRM).
With Kuala Lumpur making strides towards a safer urban environment by utilising a network of smart AI cameras, how are the other states and cities in the country faring?
Wider adoption
In Putrajaya and Penang island, there has been a similar adoption of AI-powered cameras, with the local authorities laying the groundwork for smarter surveillance systems and further integrating AI capabilities to improve public safety, monitoring, and urban management.
According to Mohd Bashir Sulaiman, the Penang Island City Council (MBPP) director of planning and development, 1,233 cameras have been deployed across the city as of June 2026.
He says that 496 of the cameras have already been integrated into the AI system under the council’s smart city initiative, while the remaining 737 are in the process of being connected.
“The rollout is being implemented in phases, focusing first on high-traffic and high-priority areas. Selected cameras have been upgraded with AI capabilities, while others remain conventional CCTV units.
“While not all cameras are AI-enabled, we have strategically deployed smart surveillance features in key locations. This phased approach allows us to evaluate system performance and ensure operational readiness before expanding further.
“By using AI analytics, the system can automatically flag unusual situations, enabling faster action from relevant response teams,” he says, further highlighting that MBPP aims to expand with an additional 726 CCTV cameras between Q3 2026 and Q3 2028 for a total of 1,959 in the city. Of these, 796 will be AI-powered.
Mohd Bashir adds that it allows real-time monitoring and computation of various aspects of city operations and incidents going on throughout Penang island.
He highlights that “the system is primarily focused on traffic monitoring and road safety management”. The Penang island deployment of AI cameras involves the use of both edge and cloud computing in video analysis.
Each computing method serves a specific purpose, with edge computing (referring to on-device processing of data) covering Automatic Number Plate Recognition (ANPR), vehicle counting, and vehicle classification.
On the other hand, the server-based cloud computing handles traffic congestion, accidents or vehicle stoppage, illegal dumping, water ponding (the accumulation of water on surfaces), the identification of abandoned objects, human behaviour, and crowd activities.
Both computational methods contribute to faster detection and response times, allowing relevant authorities to act more quickly in the interest of public safety, Mohd Bashir says.
Meanwhile, in Putrajaya, the Putrajaya Corporation (PPj) is similarly looking to expand its existing network of smart CCTV cameras. Currently, it has over 560 deployed and targets an additional 700 in the near future.
PPj president Datuk Mohd Sakeri Abdul Kadir said the cameras are monitored through the Putrajaya Command Centre (PCC), where AI is used to analyse multiple camera feeds simultaneously, helping operators identify incidents and respond more efficiently.
“AI-powered cameras support Putrajaya’s goal to become a more integrated and responsive smart city,” he says, pointing out that PCC functions to “bring together CCTV, traffic monitoring, environmental sensors and incident response”.
“In the long term, PCC aims to evolve into an Intelligent Operation Centre. The vision is to shift from reactive city management to proactive, data-driven decision-making,” he says, adding that it works to strengthen Putrajaya’s ability to manage the city.
Shared solutions
When it comes to integration and accessibility to government agencies and law enforcement, Mohd Bashir points to MBPP’s Intelligence Operations Centre (IOC), which serves as a central hub for real-time data collected from cameras across the city.
The IOC is integrated with platforms such as the Virtual Island of Penang (VIP), allowing information to be shared across systems while automated alerts notify enforcement officers and police, helping them respond more swiftly to incidents.
“We currently share access with PDRM through a dedicated interface for enforcement and investigative purposes.
“At this stage, the system is not yet integrated with other platforms or agencies. However, future integration may be considered based on operational needs and policy decisions,” Mohd Bashir says.
Similarly, Mohd Sakeri says that PCC acts as an integrated platform that various government agencies can use in real-time monitoring and incident response, similar to what is being deployed in Kuala Lumpur’s KLCCC.
“Integration is already in place with agencies such as PDRM, Bomba, Hospital Putrajaya and Civil Defence. During incidents, relevant agencies are able to access shared live CCTV feeds and situational data.
“The ‘Duplication View’ concept enables Mini PCC setups within partner agencies for synchronised decision-making. This approach improves coordination, response speed and public safety outcomes.
“The goal is to move from siloed operations to a coordinated, data-driven city management model,” he says.
When it comes to the challenges involved in deploying and maintaining a system that spans an entire city, Mohd Bashir points to the significant storage capacity required to handle the large volumes of data produced by the high-resolution cameras.
The footage and data captured also need to be transmitted to MBPP’s cloud servers for further computing, requiring reliable and high-speed connectivity to support real-time monitoring.
These factors contribute to the need for infrastructure to set up such a system and maintain it beyond the initial deployment, ensuring the system’s long-term sustainability remains a key consideration.
He says that MBPP is “continuously working to strengthen these areas to ensure optimal system performance”.
Addressing privacy concerns
How all that data is handled is a key consideration with the deployment of such systems, as highlighted by intellectual property, information technology, and privacy lawyer Foong Cheng Leong.
“Members of the public may not know what information is being collected, where it is stored, how long it is retained, who may access it, the purposes for which it may be used or the level of security applied.
“Local councils are not subject to the Personal Data Protection Act 2010 (PDPA) and may therefore not be legally required to comply with the data protection principles,” he says.
Foong further states that the right to privacy includes the right for individuals to control their own information, which may be eroded due to the information being captured and stored.
Asia Pacific University of Technology and Innovation (APU) head of forensics and cybersecurity research centre, Dr Julia Juremi, similarly stresses the need for transparency.
“I think most Malaysians understand that technology has an important role in making our cities safer. At the same time, people also want the assurance that their privacy is being respected,” she says.
Foong believes that while there is a lower expectation of privacy when an individual is in a public space, such AI-powered cameras may raise concerns as they could be used to identify individuals, track their movements, and analyse their activities over time.
Julia concurs: “If attackers gained access to this data, they could use it for identity theft, targeted scams, stalking, or even planning physical crimes by understanding people’s daily routines. If facial recognition data is involved, the risks become even greater.”
Foong also warns that these databases containing facial images, biometric information, and location histories are attractive to threat actors, as the data can then be abused for various malicious activities.
“Such information may be stolen, sold or combined with other leaked information for scams, identity fraud, surveillance and other criminal activities. Unlike passwords, biometric information cannot easily be changed once compromised,” he says.
Julia also warns that in the hands of bad actors, footage from the system could even be used for impersonation or deepfakes to enable more convincing social engineering attacks, especially when combined with other publicly available information.
Such concerns are being taken into consideration by the city councils that have implemented an AI-powered network of CCTV cameras.
Mohd Bashir says the Penang island CCTV camera platform complies with all the relevant acts and laws, including PDPA, and is deployed within a private network environment with multiple layers of cybersecurity controls in place.
“These include enterprise firewall protection, DDoS protection, network access control, encrypted communication, role-based user access (RBAC), password encryption, and video/data access protection within the Video Management System.
DDoS refers to a distributed denial-of-service attack, which is a malicious attempt to disrupt the normal traffic of a targeted server, service, or network; while RBAC refers to a security model where users are authorised to access a system based on their specific roles within an organisation.
“The supporting servers are also managed with standard security practices such as security patching, endpoint protection, and system monitoring,” Mohd Bashir explains, adding that “access to the system is strictly limited to authorised personnel”.
Meanwhile PPj’s Mohd Sakeri says that the data gathered is strictly handled in compliance with their information security management system: “Cyber-security remains a key concern, particularly when integrating systems with different technical standards and security levels.”
According to him, data collected by the system is stored in PPj-managed data centres and retained only in accordance with government operational policies.
“Security measures include encryption, access controls, audit trails and continuous cybersecurity monitoring. Access will be limited to authorised personnel and relevant agencies under defined protocols.
“The priority is to deliver public benefits while protecting individual rights and maintaining transparency.
“Access must be controlled, authorised, and purpose-driven to protect data security and public trust,” he says.
