SINGAPORE – The Traffic Police (TP) are planning to use artificial intelligence to identify traffic violations, allowing them to dish out penalties and summonses more quickly.
TP now receives about 3,000 videos monthly of alleged traffic violations submitted by the public.
But owing to the deteriorating road traffic situation, there have been more such submissions.
An AI-powered video analytics tool called Traffic Rule Analysis and Compliance Recognition (TRACER) is an example of how the police will be using AI in its operations, said Second Minister for Home Affairs Edwin Tong.
Speaking at the Police Workplan Seminar 2026 held at the Home Team Academy on May 22, he said technology is necessary for the Singapore Police Force (SPF) to stay ahead of threats with limited manpower.
Second Minister for Home Affairs Edwin Tong speaking at the Police Workplan Seminar 2026 held at the Home Team Academy on May 22.
ST PHOTO: BRIAN TEO
He said TRACER is being progressively rolled out, and can automatically identify traffic violations in footage submitted by the public, pinpointing their exact timestamp within the recordings.
He said: “This allows the officers to process cases faster, and deal with the offending drivers faster as well.
“These tools enhance the efficiency as well as the effectiveness of SPF’s investigations. Not by replacing the judgment of our officers, but really by sharpening it and freeing up the officers to focus on what only a human officer can do.”
The increase in public video submissions has caused a backlog of cases, as there are only about 10 officers who spend around 60 hours daily examining the videos.
The current version of TRACER can detect offences such as crossing double white lines, driving across road dividers, driving on the centre divider of the expressway, running red lights and failing to give way to pedestrians.
It is being developed to also detect those who fail to signal, fail to give way to approaching vehicles and fail to form up correctly.
Officers can review the key moments in the videos instead of watching the entire clip each time, allowing them to process cases faster and issue notices to offending drivers earlier.
