The Future of Autonomous Vehicles: Lessons Learned from a Recent Accident
By now, most of us are familiar with the potential of autonomous vehicles to revolutionize transportation. However, a recent accident involving General Motors’ autonomous vehicle division, Cruise, has brought to light some uncomfortable truths about the development of these vehicles and their impact on society.
In late October, one of Cruise’s vehicles struck a pedestrian who had been thrown onto the road after being hit by another car. While the autonomous vehicle initially came to a stop, it then moved an additional 20 feet to free the seriously injured pedestrian. The incident led to the suspension of Cruise’s operations in California, raising concerns about the trustworthiness of this technology. Despite this setback, if Cruise and its competitors act swiftly and demonstrate that they have learned from this incident, it does not necessarily mean that the development of autonomous taxis will be delayed by years.
This accident has shed light on several uncomfortable truths about autonomous vehicles and artificial intelligence. Firstly, the race to develop these groundbreaking technologies creates inevitable tension. On one hand, there is the atmosphere of rapid technological advancement in Silicon Valley, while on the other hand, there is the need for safety and mature processes that have been developed over years in more established markets.
Secondly, the accident highlights that deep learning, the technology behind some of the most advanced artificial intelligence systems today, is still not advanced enough to predict accidents like the one that occurred with Cruise. It may never be possible. This accident shows us that continuous learning systems are only as good as the quality of the data they are fed. Regardless of the amount of data available, it is impossible to train them on all possible situations they may encounter in the world.
In order to regain public trust, Cruise and its rivals must prove not only that their vehicles have fewer accidents than human drivers, but also that they do not make serious mistakes that a human could avoid. This is still a high standard for today’s technology to meet.
The third issue raised by this accident pertains to the regulation of autonomous vehicles. While there is much discussion about how to regulate artificial intelligence, less attention is given to who should be responsible for regulation and what influence ordinary citizens and their elected representatives should have on a technology that deeply affects their lives.
Cruise’s proactive response, including an external investigation into its technology and accident management, as well as the voluntary withdrawal of its vehicles, demonstrates their understanding of the gravity of the situation. This is an opportunity for Cruise and other companies to convince the world that this incident is a turning point in the development of autonomous vehicles.