GM Begins Public Road Testing for New Eyes-Off Driving Tech | California & Michigan

GM Accelerates Autonomous Vehicle Testing on Public Roads

General Motors (GM) is significantly expanding its testing of next-generation automated driving technology, initiating supervised public-road trials on highways in California and Michigan this week. The move, announced March 23, 2026, represents a crucial step toward the company’s goal of offering “eyes-off” driving capabilities to consumers by 2028.

More than 200 test vehicles, staffed by trained drivers capable of taking manual control, will be deployed on limited-access highways. This phase marks a transition from primarily data collection to active testing of the automated system in real-world traffic conditions. GM’s strategy hinges on leveraging a massive dataset accumulated over years of driving – exceeding one million miles across 34 states – to refine its AI driving model.

The forthcoming “eyes-off” system, initially slated for the Cadillac Escalade IQ, aims to provide a self-driving experience that doesn’t require constant driver vigilance for safety. GM intends to expand availability beyond Cadillac to encompass a broad range of its vehicle portfolio, from premium models to mainstream Chevys. This ambitious plan relies on a new centralized computing architecture designed to consolidate vehicle intelligence, eliminating the need to rebuild the system for each vehicle type.

GM’s approach to autonomous driving is built on a foundation of real-world data and extensive simulation. The company highlights the more than 800 million miles driven by customers using Super Cruise, its existing hands-free driver-assistance system, without any reported crashes attributed to the technology. This substantial safety record, combined with over 5 million fully autonomous miles logged by Cruise in complex urban environments, provides a robust baseline for the development of more advanced automated features.

To complement real-world testing, GM utilizes a sophisticated simulation environment capable of replicating roughly 100 years of human driving each day. This allows engineers to rapidly test and refine the AI driving model in a wide range of scenarios, accelerating the development process and enhancing system robustness. Data gathered during the current phase of public-road testing will be fed directly back into this development cycle, further improving the system’s performance.

The move by GM reflects a broader industry trend toward scaling autonomous technology. While early efforts focused on isolated robotaxi deployments, companies are increasingly prioritizing the integration of automated features into personal vehicles. GM’s strategy of deploying a common architecture across its entire fleet aims to leverage manufacturing scale and reduce the costs associated with developing and maintaining multiple autonomous systems. The company’s engineering blog, recently launched, will provide ongoing updates on these technological advancements.

Looking ahead, the success of GM’s autonomous driving initiative will depend on continued regulatory progress and the operational resilience of its technology. The company’s ability to navigate these challenges will determine whether it can successfully establish a scalable autonomous driving paradigm and deliver on its promise of “eyes-off” driving by 2028.