Dhaka’s new AI traffic cameras are successfully taming the capital’s chaotic roads, but the high-tech crackdown is sparking a quiet debate over people’s privacy and security.
This technology automatically detects violations ranging from running red lights and wrong-way driving to lane infractions, illegal parking, and stopping beyond designated stop lines. However, while its efficacy is widely praised, critical concerns regarding data protection, security, and citizen privacy remain unaddressed.
In April this year, the Dhaka Metropolitan Police (DMP) officially launched the “AI-Based Road Transport Act Violation Detection System.”
The first phase deployed 25 cameras along the corridor from Hotel InterContinental to Jahangir Gate, before expanding to key intersections such as Shahbagh, Bangla Motor, Karwan Bazar, Bijoy Sarani, and Airport Road.
According to the DMP, the pilot programme processed approximately 1,000 traffic cases and over 5,000 video clips within its first 10 days.
Building on this momentum, authorities plan to scale the technology to hundreds of other major intersections across the capital.
Police officials claim the system has already fostered better traffic discipline, noting a decrease in red-light violations and an increase in drivers respecting stop lines. Furthermore, it has reduced the need for manual human intervention.
Yet, despite these promising results, a pivotal question persists: even if the technology excels at enforcement, is it operating within an adequate framework of security and accountability?
Lack of transparency, international security concerns
There are valid reasons for these doubts. While there has been extensive debate surrounding the effectiveness of the technology, discussion regarding its source, infrastructure, and cyber security remains relatively limited.
The DMP has not yet disclosed detailed information about the camera manufacturer, the software used, the network architecture, or the data storage system.
While many countries keep such details confidential for security reasons, experts argue that maintaining a minimum level of transparency and accountability is equally essential.
This concern has been further heightened by recent international events.
In February this year, a high-profile attack on Iran’s Supreme Leader, Ayatollah Ali Khamenei, led to international reports claiming that Israeli intelligence agencies had been hacking into a vast network of traffic and surveillance cameras in Tehran for years. According to these reports, the compromised data was used to analyse the movements, security routines, and communication patterns of key figures, enabling targeted attacks.
British and US media outlets also reported that numerous traffic cameras in Tehran had been compromised over an extended period, with video data being routed to foreign servers.
By leveraging artificial intelligence, a vast amount of video footage was analysed to map individuals’ ‘patterns of life’ and daily movements.
Following these disclosures, a fresh global debate has emerged regarding the security of surveillance infrastructure.
Some reports claim that Russia has also re-evaluated its use of certain surveillance systems to ensure the safety of high-profile individuals.
Consequently, modern surveillance technology has firmly transitioned into the realm of national security.
Important questions for Bangladesh
The rapid advancement of AI-based video analytics has intensified global security debates. At the centre of this is Israel-based defence tech firm Iris Labs, which recently secured a $60 million investment for a platform that aggregates data from security cameras, drones, and social media into actionable intelligence.
While unlinked to the Tehran hacks, it highlights how quickly video analytics are being integrated into intelligence operations.
This context raises critical unanswered questions for Bangladesh.
No official information has been disclosed regarding the manufacturers of Dhaka’s surveillance cameras, where the software is hosted, whether data is stored domestically or on foreign cloud servers, or who manages cybersecurity.
Unconfirmed reports hint at brands like Tyco, Hikvision, and Dahua, but official clarity is lacking.
Globally, surveillance infrastructure is increasingly treated as a national security matter rather than a purely technical asset.
On 25 November 2022, the US Federal Communications Commission (FCC) issued a statement introducing new rules restricting the approval and marketing of equipment from Chinese firms Hikvision and Dahua over national security concerns.
Similarly, an Associated Press report in February 2023 revealed that the Australian government decided to remove hundreds of cameras from these same manufacturers across defence and government facilities following a comprehensive security audit.
Although the accused companies deny these allegations, these developments show that modern surveillance technology is now deeply intertwined with national security, data sovereignty, cyber security, and geopolitical competition.
Experts’ concerns
Against this backdrop, technology experts are raising serious questions about the security and accountability of Bangladesh’s AI-based traffic camera system.
According to Anindya Iqbal, a professor of Computer Science and Engineering at the Bangladesh University of Engineering and Technology (BUET), such technology-driven surveillance systems can become a major source of concern for citizens if they lack adequate transparency and independent oversight.
“The responsibility does not end with the deployment of technology; rather, ensuring the accuracy and integrity of that technology is equally important,” Dr Iqbal noted.
He emphasised that the biggest question surrounds the system’s credibility, arguing that authorities must be far more transparent about how the system operates, where the data is stored, who has access to it, and whether it is being used for any purpose other than its intended function.
At the same time, he urged stakeholders to treat cyber security with the utmost seriousness.
He warned that Bangladesh’s past experiences demonstrate that the risk of major cyber attacks can never be entirely ruled out, meaning the potential leakage of citizens’ location data or other sensitive information remains a real threat.
In this context, he cited the 2016 Bangladesh Bank reserve heist as a cautionary tale. At that time, cyber criminals managed to steal approximately $81 million by exploiting weak cyber security protocols, outdated technology, and vulnerabilities in network security. Experts argue that this incident highlights how even a minor flaw in critical digital infrastructure can lead to national-level consequences.
Sharing a similar perspective, Aminul Bari Shuvro, chief strategy officer of Synesis IT, a leading Bangladeshi software company, believes that while AI-based traffic cameras play a positive role in improving traffic discipline, data protection and civil liberties cannot be overlooked.
He observed that while artificial intelligence and digital transformation are widely discussed within the country, relatively little attention is paid to data security and cyber risks.
He stressed the importance of establishing clarity and accountability regarding where, how, and by whom citizens’ information is being handled.
“The public owns their own information,” Shuvro said. “Therefore, if an institution collects citizens’ data, it must ensure public consent and maintain absolute transparency in how that information is managed.”
He added that building public trust in technology-based surveillance systems will be difficult without a robust legal and ethical framework.
Security analyst Colonel (Retired) Kazi Sharif views the issue through a broader national security lens.
He argues that AI-based traffic cameras are not merely a tool for traffic management; they are intrinsically linked to national security. In the modern era, the nature of warfare has shifted, with technological, cyber, and electronic warfare becoming increasingly prominent, he said.
Consequently, he concludes, large-scale surveillance infrastructure must be rigorously evaluated as a vital component of national security.
Experts’ proposed solutions
In light of these challenges, Dr Iqbal believes that regular vulnerability assessments, penetration testing, and independent third-party technical audits are essential for AI-based traffic surveillance systems.
International standards strongly echo this approach. For instance, the guidelines established by the UK’s Biometrics and Surveillance Camera Commissioner’s Office highlight independent security audits and third-party certification as vital safeguards for large-scale public surveillance infrastructure.
Shuvro emphasises the absolute necessity of transparency and accountability regarding where, how, and by whom citizens’ data is used.
“The public owns their own information,” he notes. Consequently, he argues that public consent, transparency, and accountability must be legally guaranteed whenever citizens’ data is collected, while also calling for the swift implementation of a robust personal data protection act.
Colonel Sharif, meanwhile, believes the issue demands urgent debate at the highest policy-making levels, including within parliament. He warns that while widespread camera networks can be highly effective in curbing crime, they also pose a severe threat to personal privacy and civil liberties if left unchecked or misused.
What DMP says
The DMP does not fully agree with the concerns raised by experts. It maintains that the currently operational AI-based traffic camera system is used solely for enforcing traffic regulations and does not directly collect citizens’ personal data.
Mohammad Nurus Samad Bapi, senior system analyst at the DMP’s Traffic Department, told TIMES, “Our role is strictly limited to issuing prosecutions based on the vehicle registration plate once an offence has been identified.”
He added that all data stored within the system is encrypted, and the entire infrastructure operates over an internal network and a virtual private network (VPN).
