Determining Liability When AI, Robots Cause Accidents
- By DSAITrends editors
- October 07, 2020
The laws in Singapore are currently not able to effectively assign liability when it comes to losses or harm suffered in accidents where AI and robotics are involved, say panelists at the third TechLaw.Fest forum last week.
Determining liability
“The unique ability of autonomous robots and AI systems to operate independently without any human involvement muddies the waters of liability,” said Charles Lim at a webinar reported by broadsheet The Straits Times.
“There are multiple factors (in play) such as the AI system's underlying software code, the data it was trained on, and the external environment the system is deployed in,” he said. Lim is a lawyer and co-chair of the Singapore Academy of Law’s Subcommittee on Robotics and Artificial Intelligence.
Confusing the picture would be whether an accident could stem from a latent bug, introduced by a recent software update, or influenced by a human’s decisions in the events leading up to the accident. Retracing the precise sequence of events leading up to the accident to prove liability can be complex and cost, says Lim, who spoke in his personal capacity at the webinar.
In September, the 11-member subcommittee published a report (pdf) “Report on the Attribution of Civil Liability for Accidents Involving Autonomous Cars” that discussed the attribution of civil liability for accidents involving autonomous cars (AV).
Singapore is investing heavily in AI as it seeks to lead the region in AI experimentation and to establish itself as a premier destination for global AI investment. In 2017, the National Research Foundation (NRF) launched AI Singapore (AISG), an SGD150 million (~USD109 million) national program to deepen its AI capabilities for the future digital economy.
Last year, the Singapore Land Transport Authority announced that the entire western part of the island state will eventually become a testing ground with more than 1,000km of public roads opened for the testing of self-driving vehicles. This represents a substantial set up from previous 70km of public roads available for the testing of AVs.
Some of the firms testing AVs right now are ST Engineering, listed transport giant ComfortDelGro, and MIT spin-off startup nuTonomy.
Photo credit: iStockphoto/monkeybusinessimages