Sign In

Communications of the ACM

ACM TechNews

Hackers Can Trick a Tesla into Accelerating By 50 Miles Per Hour

View as: Print Mobile App Share:
A Tesla on the road.

Aadversarial machine learning can thwart autonomous driving systems, according to researchers at McAfee.

Credit: Tesla

Researchers at McAfee have demonstrated that adversarial machine learning can thwart autonomous driving systems, presenting a security challenge for the emerging technology.

Tesla's MobilEye EyeQ3 camera systems read speed limit signs and send that information into autonomous driving systems. The McAfee researchers placed a tiny and nearly imperceptible sticker on a speed limit sign, causing the camera to read the sign as 85 rather than 35.

Test results revealed that both the 2016 Tesla Model X and the 2016 Tesla Model S sped up 50 miles per hour as a result of the adversarial incident.

Said McAfee researcher Steve Povolny, "If we are not very prescient about what the attacks are and very careful about how the systems are designed, you then have a rolling fleet of interconnected computers which are one of the most impactful and enticing attack surfaces out there."

From Technology Review
View Full Article


Abstracts Copyright © 2020 SmithBucklin, Washington, DC, USA


No entries found