A police officer supervises traffic violations through an AI camera system installed on the HCM City-Long Thành-Dầu Giây Expressway. VNA/VNS Photo

Transportation expert Vũ Hoàng Chung speaks to the Kinh Tế & Đô Thị (Economic & Urban Affairs) newspaper about the application of Artificial Intelligence (AI) cameras in traffic supervision and their roles in assisting police officers.

How do you evaluate the application of AI cameras for traffic violation surveillance and enforcement, which the Traffic Police Department is currently piloting?

The application of modern information technology, and especially AI cameras, to traffic and all aspects of social life is an inevitable trend for any country.

Digital transformation, leading to a digital society and digital citizens, is the right path forward and will bring comprehensive efficiency. The use of AI cameras for traffic surveillance and enforcement is therefore necessary and has been effective.

AI cameras don't just serve to enforce traffic violations; they also help with monitoring and assisting in traffic regulation, accident resolution and maintaining security and order.

If data is well connected and fully utilised, it can bring many other benefits to society.

Some people believe that with an AI camera system, the role of the traffic police will become more blurred, or they won't even need to be on the streets anymore. Do you agree?

I don't think so. The nature of an AI camera is to be a tool that assists the traffic police. The role of the traffic police in operating it and enforcing the law is irreplaceable.

Cameras can help collect data and AI can provide optimal solutions for situations, but the final decision is still made by people.

Traffic police officers still have to be present daily to maintain order and traffic safety.

However, there are some stages in the process of handling traffic violations that have previously been seen as complex and subject to conflicting opinions, with people questioning their transparency, especially when violations are handled on the road.

This is precisely why we need an automated AI camera system to record violations, issue tickets and send notifications to citizens.

This will minimise contact between traffic police and citizens, making the process more transparent and fair.

It's clear that AI cameras don't blur the role of traffic police, but serve as a tool to enhance their role and effectiveness.

For example, in traffic diversion and regulation, data obtained from cameras will help detect congestion early and from a distance, allowing for the discovery of alternative routes to alleviate pressure.

When a violation occurs, people will receive an automatic notification to pay the fine, eliminating opportunities for them to plead for leniency or seek intervention.

This is very important because it encourages citizens and businesses to raise their awareness and voluntarily comply with the law.

When they can't ask for favours or get out of fines, the phenomenon of flouting or circumventing the law will disappear.

How do you explain the opinion that traffic police will no longer have to stop and apprehend violators on the streets, especially given the current high number of violations?

If AI cameras could truly help traffic police avoid stopping violators on the streets, that would be wonderful.

In many cases, people fearing a fine are willing to offer bribes or flee when they see a traffic police officer.

If direct enforcement is eliminated, these risks could gradually be eradicated, while violators are still handled fairly and transparently. That's a great outcome.

However, I believe the idea that traffic police will no longer have to be on the streets to handle violations is incorrect.

Even in the most developed countries, police still have to patrol and handle violations on the road. This is because some violations require immediate action to prevent the risk of accidents, such as speeding or drunk driving.

If traffic police aren't stationed to check, how can they detect if a driver has violated alcohol limits?

Traffic police will thus still be needed to patrol and handle violations on the road.

The only difference is that with AI cameras, direct enforcement will be minimised.

In reality, the traffic police force does not have enough personnel to directly handle all violations.

Detecting violations through cameras is necessary to ensure all violations are addressed.

This is especially true for violations by businesses and individuals using commercial vehicles.

With a comprehensive system of electronic scales and AI cameras, violations like overloading, oversized loads or indiscriminate stops for picking up/dropping off passengers or goods will be resolved completely.

I believe that traffic police won't need to be on the streets to apprehend those who commit these types of violations, and can still handle them with an AI camera system.

What conditions are needed, in your opinion, to build and operate an effective AI camera system for traffic surveillance and enforcement?

First, there must be a good AI camera system with full funding that ensures connectivity with population data from all other sectors and fields.

We are not installing an AI camera system to be used for just five to 10 years, but for much longer, so it needs to be given the proper attention.

Many cities around the world are currently renting IT infrastructure from specialised businesses. The State rents the system, businesses invest in and upgrade the system, and citizens benefit. That's a direction worth considering.

Next, we need a skilled workforce. Previously, traffic police were trained to operate in a real-world environment. Today, their tasks and enforcement tools have changed, becoming increasingly modernised and digitised. Traffic police must therefore also be trained to become 'digital officers'.

The construction of the national population data system and transport infrastructure needs to be accelerated and prioritised. Data is the foundation for all digital tools to operate effectively.

Without population data to investigate, AI cameras would be no different from regular cameras and would not help solve complex traffic problems.

For Hà Nội, do you think the implementation of AI cameras for traffic surveillance and enforcement will be effective?

A large city like Hà Nội has an even more urgent need for tools like AI traffic cameras.

Hà Nội is currently facing a complex problem of traffic congestion and violations. The use of AI cameras will mitigate this situation and significantly contribute to building a better traffic culture among the people of the capital.

On the other hand, fair and transparent enforcement also helps raise the image of traffic police officers to become more friendly and approachable for the public.

This is especially important for a city with nearly 10 million people, eight million vehicles and a leading political, economic, social and cultural role in the country.

Hà Nội needs to be at the forefront of such an important digital transformation. VNS