01:59 10-05-2026
Autopilot that almost writes itself: GM lifts the lid on AI's role
CEO Mary Barra reveals AI generates almost 90% of GM autonomy code, with simulators modeling 100 years of human driving every day.
General Motors is leaning ever harder on artificial intelligence in the development of its self-driving systems. Speaking at the company's first-quarter 2026 earnings presentation, CEO Mary Barra said that nearly 90% of the code being written by GM's autonomy team is now generated by AI.
That code, GM stresses, is vetted in a digital environment. According to Barra, the system is stress-tested in a simulator capable of modeling roughly 100 years of human driving every single day. The aim is to accelerate validation of complex road scenarios without having to wait for them to occur in real-world traffic.
The new technology is intended for GM's next generation of automated driving. The «eyes-off» system is set to debut in 2028 on the electric Cadillac Escalade IQ. Today's Super Cruise allows hands-free driving but still requires the driver to keep their eyes on the road; the future setup is meant to let drivers look away from the road on compatible highways across North America.
GM is betting on more than cameras alone. Unlike Tesla's approach, the upcoming system will rely on a broader sensor suite: lidar, radar and cameras. Under SAE classification, that puts it at Level 3, meaning the car can take over certain tasks under defined conditions, while the driver remains part of the system and must be ready to retake control.
The new autopilot will also tie into GM's centralized computing architecture. The company promises a step-change in capability: 10 times more bandwidth for over-the-air updates, 1,000 times more internal bandwidth, and up to 35 times more AI performance for self-driving features.
For now, GM is gathering data through Super Cruise. The company says customers have logged more than 1 billion hands-free miles on North American roads, with around 750,000 vehicles equipped with the system on the road.