As soon as a vehicle stops at the traffic lights at Wong Sawang intersection, a loud alarm goes off, followed by a pre-recorded announcement in Thai, saying "Please don't obstruct the zebra crossing. Thank you."
The alarm and announcement will keep playing, until all motorists who are waiting for the red line to turn green move their vehicles from the pedestrian crossing.
This is artificial intelligence (AI) in action, which City Hall is marshalling to a civic cause.
In an effort to improve pedestrian safety across the capital, City Hall has installed several cameras fitted with sensors at two major traffic junctions in Bangkok: the Wong Sawang and Arun Amarin intersections.
Not only will the cameras act as a deterrent, they could also take a picture of both the offence and the offender's image, says the director of the Traffic Technology System Division at the Bangkok Metropolitan Administration (BMA), Rapeepong Morakul.
"The alarm and the warning announcement are loud enough to embarrass offending motorists," he said, expressing his confidence the AI-assisted cameras will be able to reduce traffic violations.
That said, the pictures will only serve as a warning for the time being, as authorities have no plan to prosecute offenders caught by the AI-assisted cameras during the trial period, he said.
Those who intentionally run a red light, however, will be prosecuted if they are caught by the cameras.
Authorities decided to deploy the cameras to stop motorcycles and cars from obstructing zebra crossings for the benefit of the capital's pedestrians.
Such violations remain rampant across the city despite the rules in place against such an offence.
Initiated by the previous Bangkok governor, Aswin Kwanmuang, this project was finally implemented after governor Chadchart Sittipunt signed a 106-million-baht contract with Samart Comtech Co on Aug 31 last year, hiring the company to install AI-assisted cameras at 20 major crossings in city, said Mr Rapeepong.
Eighteen more locations will be fitted with such cameras, and those include Huai Khwang, Suan Ruen, Nida, Kan Kheha (Thi Din), Si Ayutthaya, Ratchawat, Ramkhamhaeng, Chao Khun Thahan, Ratchaprasong, Sala Daeng, Henri Dunant, Ratchayothin, Phang Muang, Sakae Ngam, Pharam 9, Wang Doem, Bukkhalo and Tha Din Daeng intersections.
It takes roughly about two months to install cameras at each location, as the cameras must be placed high enough so they can capture a bird's eye view.
The system will be further developed so that all the data from the cameras can be used to control the traffic lights below, he said.