Recent reports of two near misses involving autonomous vehicles have drawn public attention to the safety of self-driving technology. According to media reports, two autonomous vehicles in Punggol narrowly avoided hitting a pedestrian and another vehicle respectively last Thursday (5 November). Members of the public subsequently shared videos online showing autonomous vehicles nearly colliding with other road users and apparently violating traffic rules. Some netizens said they had repeatedly seen autonomous vehicles stop abruptly on the road, including braking suddenly midway through a yellow box junction.

Responding to queries from Lianhe Zaobao, a Land Transport Authority (LTA) spokesperson confirmed that the authority was aware of the situation and was reviewing the incidents with the autonomous vehicle operators concerned to establish what happened and identify any necessary improvements.

Beyond concerns about the safety of self-driving technology, the public is also keen to know who would be held liable in an accident involving an autonomous vehicle. LTA said all autonomous vehicles currently on the roads are governed by the regulatory sandbox framework under the Road Traffic (Autonomous Motor Vehicles) Rules. They must comply with all traffic rules, as well as LTA’s safety and operational requirements.

LTA has also worked with the Traffic Police to develop a protocol for responding to and managing incidents involving autonomous vehicles. All authorised autonomous vehicle operators must maintain liability insurance covering third-party deaths, bodily injuries and property damage caused by or arising from the use of these vehicles, including claims involving passengers.

People’s concerns about this new technology are not unfounded. History shows that the harm caused by an emerging technology can sometimes become apparent before its benefits. Before atomic energy could light up entire cities, it destroyed Hiroshima and Nagasaki. Artificial intelligence has immense capabilities, yet has prompted warnings that it could threaten humanity’s survival.

Though on a different scale, the enormous potential of self-driving technology has been accompanied, during trials, by concerns ranging from road safety to liability for accident compensation. But the fact that new technologies drive human progress means that embracing them is the right approach. Addressing the public’s various concerns about autonomous vehicles must therefore be a priority.

Self-driving technology is gradually being deployed around the world. The US is one of the most mature markets for its commercial use. Waymo already offers paid, fully driverless robotaxi services in 15 metropolitan areas, including the San Francisco Bay Area, Los Angeles, Miami, Dallas and Houston.

China’s Apollo Go, Pony.ai and WeRide also operate in major cities such as Guangzhou, Shenzhen and Beijing. Japan, which has taken a more cautious approach, plans to introduce a unified certification system from December for autonomous taxi and bus services using vehicles without steering wheels or pedals. Trials are currently under way in several districts of Tokyo. Europe is relatively far behind: the UK and Germany are still conducting trials, while the EU has established type-approval rules for the automated driving systems of fully autonomous vehicles.

US data on autonomous vehicle safety is relatively comprehensive and verifiable. An analysis by the Insurance Institute for Highway Safety of data covering about 50 million miles travelled between 2021 and 2024 found that Waymo’s rate of police-reportable crashes was 68% lower than that of human drivers in the same areas, while its rate of injury crashes was 81% lower. It has recorded no fatal crashes to date.

This is largely attributable to the absence of risks associated with human drivers, such as fatigue, distraction or drink-driving. A new technology still requires a longer observation period and more data before the evidence becomes sufficiently convincing. Even so, such findings offer a better basis for assessing the safety of self-driving technology than public perceptions shaped by a handful of near misses.

Every city has its own road culture, and the unspoken understanding among drivers, motorcyclists, cyclists and pedestrians varies from place to place. While overseas experience and data provide useful references, we must still rely on our own practical experience to determine how best to deploy self-driving technology.

The incidents in Punggol are therefore an opportunity to identify and resolve problems — an effort that requires participation across society. As new technology arrives, the public should keep an open mind rather than dismiss it hastily. Of course, the safety of self-driving technology must remain the foremost consideration, with no room for compromise.

At the same time, the authorities have pledged to maintain transparency. From next month, they will publish quarterly reports on the deployment of autonomous vehicles, including the size of the fleet in service, passenger experience and safety performance. This will give the public access to actual data, allowing a more objective assessment of self-driving technology’s performance.

As with the arrival of any previous new technology, resisting it outright would clearly be unwise. Yet Singapore’s densely populated urban environment calls for a cautious, measured approach, with wider deployment only after the technology’s safety has been established. Given the global trajectory of technological development, this may be the only viable course.