Academics say that beyond enforcing the law, deploying artificial intelligence can help to identify behaviour and conditions that endanger road users.
For instance, it can help the authorities detect risky driving behaviour and identify blackspots before accidents happen, they told FMT.
They said that employing AI to analyse traffic patterns, monitor driver behaviour and flag high-risk locations will enable earlier intervention that may even help to save lives.
However, the two academics also cautioned that such a system would require access to sufficient local data, inter-agency cooperation, and human oversight to be effective.
These remarks by Yuen Choon Wah of Universiti Malaya and Sunway University’s Hadyan Hafizh are in response to the growing interest in how technology can be deployed to improve road safety following recent cases of fatal road accidents.
Yuen, who heads his varsity’s centre for transportation research, said AI can be used to identify repeat offenders, including motorists with a history of traffic violations or unpaid summonses while providing researchers with better insights into accident hotspots.
Such analysis, he added, can also highlight deeper issues related to road design and infrastructure.
Yuen noted that AI already has the potential applications that can be used to monitor traffic and enforcement, factors that can ultimately improve road safety.
“Once enforcement improves, it can help reduce traffic crashes,” he said.
However, he also stressed that AI systems cannot operate in isolation.
“We cannot rely 100% on AI output. It still has to be verified by professional engineers,” he said, underscoring the need for human oversight.
Yuen said there still are challenges that have to be overcome to ensure AI is effective in ensuring road safety. One of these is the limited access to usable data from local authorities, transport operators, and private companies.
“That is one of our biggest challenges. Without data, we cannot do much at scale,” he said, adding that stronger data-sharing mechanisms are essential.
Hadyan said discussions on AI and road safety should not be limited to surveillance and issuing of summonses.
He said AI has the potential to play a wider role. “For instance, it can be used to detect early warning signs of accidents, such as sudden braking patterns, dangerously merging into other lanes, blindspot conflicts, poor traffic light timing, potholes, and other road hazards,” he added.
“Many dangerous situations occur before an actual crash happens,” said the senior lecturer at the university’s school of computing and artificial intelligence.
Hadyan said AI systems could be deployed in high-risk areas such as school zones, motorcycle-heavy corridors, busy junctions and known highway blackspots, using data from cameras and sensors.
This, he said, could allow authorities to issue real-time warnings, adjust traffic signals, alert emergency responders or intensify enforcement in emerging risk zones.
“The most effective use of AI is not (to come up with) a single solution, but a combination of approaches working together to influence behaviour, support early intervention and improve the overall road environment,” he said.
Both researchers said any deployment should begin on a small scale before being expanded for wider use, focusing first on high-risk corridors and known blackspots.
They also stressed that AI should not be seen as a silver bullet, but rather a supporting tool that could help shift Malaysia’s road safety approach from reacting to crashes to preventing them. - FMT
No comments:
Post a Comment
Note: Only a member of this blog may post a comment.