General Motors is signaling a big shift in how cars drive themselves. At its GM Forward event, the company revealed that by 2028, it plans to bring an “eyes-off, hands-off” driving mode to production vehicles — starting with the Cadillac Escalade IQ electric SUV. This isn’t just a more powerful cruise control system; it’s a step into Level 3 autonomy, where the car takes over key driving functions under set conditions, allowing the human driver to look away from the road.

How the New Eyes-Off System Works

GM’s current Super Cruise system already lets drivers remove their hands from the steering wheel on highways that have been specially mapped for it. But even today, that system requires drivers to keep their eyes on the road: a built-in driver-attention monitoring camera watches head and eye position, ready to alert drivers if they’re not paying attention.

The upcoming system changes the rules significantly by allowing drivers to look away, read, check messages, or otherwise divert their visual focus — all while the car handles steering, braking, and lane keeping. [1]

To make this possible, GM is using a redundant sensor architecture: LiDAR, radar, and cameras all work together (sensor fusion) so the car has a more complete and reliable understanding of its surroundings. This isn’t just about seeing what’s ahead — it’s about validating that information across multiple technologies so that the car stays aware, even if one sensor type struggles in certain conditions.

When the eyes-off mode is active, GM makes its presence obvious: turquoise lighting will illuminate across the dashboard and on the exterior mirrors to alert both the driver and other road users that the system is in control.

Under the hood, GM is also revamping its computing architecture. In 2028, the company plans to roll out a centralized, high-throughput platform that unites propulsion, steering, braking, safety, and infotainment — designed for heavy AI workloads. This platform is said to deliver dramatically more processing power than GM’s current systems, with tenfold over-the-air update capacity and up to 35x the AI performance.

GM isn’t just inventing this out of thin air: the system builds on its decade-plus of experience with Super Cruise. The company has already mapped ~600,000 miles of highway for Super Cruise in North America and claims over 700 million hands-free miles driven without a reported crash attributable to the system.  GM is also leveraging expertise from its Cruise division — millions of driverless miles in simulation and testing environments feed into the new system’s AI and validation pipelines.

What This Means for the Future of Driving?

GM’s move reflects a broader realignment in its autonomy strategy. In recent years, the company has reduced its focus on robotaxi services — scaling back Cruise’s external operations and folding its technology and engineering talent into building more advanced driver assistance systems (ADAS) for consumer vehicles. [2] That shift makes sense from both a business and technological standpoint: creating a mass-market, production-vehicle autonomy system could be more efficient and safer than trying to deploy a fully driverless fleet from day one.

From a safety and trust perspective, GM’s approach is both conservative and forward-thinking. The use of redundant sensors helps manage risk, and the decision to roll out the feature first on mapped highways reduces complexity. [1] But handling real-world edge cases — unusual weather, unexpected obstacles, sensor failures — will be a major challenge. The simulation frameworks and AI models that come out of Cruise work will likely be critical here.

This new system also raises important questions about regulation, liability, and insurance. If a driver is legitimately not looking at the road, who is responsible in the event of an accident? GM hasn’t fully detailed its legal strategy, but the move hints at a future where automakers, insurers, and regulators will have to rethink how they classify driving and driving responsibility.

There’s also a big consumer behavior shift baked in. When eyes-off driving becomes available, how will drivers use it? During long highway drives, some will probably use it to relax, glance at messages, or catch up on emails. But GM will need to educate drivers on the limitations — when the system is available, when it isn’t, and when it might hand control back. That’s a delicate balance: make people too cautious and they won’t trust it; make it feel too permissive and you risk misuse.

Another layer is GM’s plan to pair this autonomy with conversational AI. The company announced that beginning in 2026, its vehicles will integrate Google Gemini to let drivers talk naturally to their car.  This AI assistant could help with route planning, drafting messages, or even explaining driving features. Over time, GM expects to roll out its own in-house AI, tuned to each driver’s preferences and the data collected from the car itself.  That means in the future, your car might not just drive for you — it might talk with you in a way that feels much more personal than today’s voice assistants.

Strategically, it’s also a signal that GM believes in autonomy at the personal-vehicle level, not just in shared, commercial, or fleet use cases. By combining ADAS and autonomous tech with a cloud-connected, software-updatable vehicle platform, GM’s laying the foundation for a new generation of smart, software-defined cars. That could position it well against competitors: traditional automakers, EV players, and even newer mobility-focused entrants.

There are real technical challenges, though. Rolling out a system that reliably supports “eyes-off” driving will require not just sensors and compute, but trusted fallback mechanisms. What happens when conditions change, map data is outdated, or unexpected events occur? GM’s reliance on simulation and real-world driving data (including Cruise’s millions of miles) will be key, but so will be how transparently the company communicates system limitations to drivers.

On the infrastructure side, even though GM has mapped hundreds of thousands of miles, the dependency on pre-mapped highways could limit how widely or quickly the system is adopted. Expanding that mapped network — especially in densely populated or rural regions — will be expensive and time-consuming. Meanwhile, legislation will likely need to catch up: what states will permit “eyes-off” driving, and under what rules or liability frameworks?

Finally, the human factor may be the wild card. Will drivers feel safe enough to genuinely take their eyes off the road? Or will they treat it like an overpowered cruise control and only glance occasionally? Will they trust the system in bad weather, heavy traffic, or construction zones? And how will behavior change over time as more drivers gain access to these capabilities?

By building out “eyes-off” driving in a carefully controlled, sensor-redundant way, GM is staking a claim on near-term personal autonomy. It’s not promising fully driverless cars tomorrow. But by layering advanced sensing, AI, and software-defined hardware, GM is paving the road for a future where cars are more than machines — they’re intelligent partners, capable of taking over when we want them to, while keeping us safe, informed, and connected.

Sources:

[1]: https://techcrunch.com/2025/10/22/gm-to-introduce-eyes-off-hands-off-driving-system-in-2028

[2]: https://news.gm.com/home.detail.html/Pages/topic/us/en/2025/feb/0225-supercruise.html

Reference:

https://www.autoblog.com/news/gm-super-cruise-to-allow-eyes-off-driving-by-2028