In a recent The Tech Advocate blog, the issue of trusting self-driving cars with our children’s safety was discussed. Imagine placing your seven-year-old child on a self-driving school bus. Would you? Researchers in China tried to answer this question. Researchers at Tianjin University and Chang’an University surveyed 499 people in the City of Tianjin to find out what level of risk would be acceptable for them when it comes self-driving cars, versus human-driven cars.[1] Their research results showed that the public would only accept this new technology if can prove to be approximately four to five times as safe as human-driven vehicles.
So, can artificial intelligence (AI) technology and the automotive industry deliver this level of safety? The goal is to make self-driving cars “safer” than human-driven vehicles. Many vehicles already have Advanced Driver Assistance Systems (ADAS) that keep drivers alerted to constantly-changing scenarios on the road. AI, machine learning and computer vision make a range of ADAS features possible, which will be part of self-driving vehicles and is currently making human driving much safer.
These features include lane departure warning, forward collision warning, rear-end collision avoidance, blind spot detection, pedestrian detection, vehicle detection, and speed limit detection. ADAS rely on sensors and actuators that communicate over localized networks. These systems are providing the foundation for future autonomous cars. This will eventually make all autonomous vehicles connected to each other. What cyber vulnerabilities does this create? Connected AI vehicles, like all Internet of Things (IoT) devices, are exposed to cyber security threats.
In 2015, cyber researchers Charlie Miller and Chris Valasek hacked a Grand Jeep Cherokee and interfered with the Jeep’s steering, braking, and other vital vehicle functions.[2] This hack was accomplished through the vehicle’s “smart system.”
A recent McKinsey survey on AI connected vehicles showed that almost half of the survey’s respondents said they distrust the computers that control autonomous vehicles and 38 percent “feared hacking.” But, more than half of the respondents indicated they would feel secure to travel in a self-driving vehicle if their concerns were addressed.
Fortunately for the industry, ADAS technologies already abide by high safety requirements. These safety requirements must, and will increase as the level of autonomy increases. Many future ADAS technologies will be rated at Automotive Safety Integrity Level D.[3] This is an ISO classification that applies to components or systems where a malfunction poses the risk of injury or death.
The automotive industry appears confident that every measure is being taken to ensure that future self-driving vehicles will be safe. Companies such as Ford, GM, Mercedes, Apple, Intel, and Delphi have all contributed major technical and financial investments to make self-driving vehicles a safe reality. The US military is aggressively researching and testing self-driving vehicles, yet personal autonomous cars are still a few years off. Once AI self-driving vehicles become prevalent, and can prove safety to its occupants, the general public will become confident in putting their young children on self-driving vehicles.
Red Sky Alliance collects and analyzes automotive cyber technology for our members.
About Wapack Labs
Wapack Labs is located in New Boston, NH. We are a Cyber Threat Analysis and Intelligence organization. For questions or comments regarding this report, please contact the lab directly by at 1-844-492-7225, or feedback@wapacklabs.com.
[1] https://www.thetechedvocate.org/would-you-let-your-self-driving-car-take-your-kids-to-school/
[2] https://resources.infosecinstitute.com/the-nightmare-of-car-hacking/#gref
[3] http://www.ni.com/en-us/innovations/white-papers/11/what-is-the-iso-26262-functional-safety-standard-.html
Comments