There’s a remarkable amount of misinformation out there about automated driving, and it’s constantly blurring the line between today’s driver-assistance systems (ADAS) and cars that can actually drive themselves. You have to know what these systems can and can’t do as we head into a future where cars are taking on more of the driving.
Key Takeaways
- Automated driving isn’t one single thing, it’s a whole range of tech, from basic aids to full self-driving, and each level has different rules for where it can operate and how much you need to pay attention.
- Level 2 systems, like the adaptive cruise with lane centering you might have now, help out a lot but you absolutely must keep your hands on the wheel and your eyes on the road.
- Real Level 3 systems, like what Mercedes-Benz is testing in Nevada, let you go “eyes-off” in very specific situations, but you still have to be ready to take back control when the car tells you to.
- The jump from Level 2 to Level 3 is a big one for liability, shifting the responsibility for monitoring the road from you to the car’s system (with you as the required backup).
- Don’t hold your breath for Level 5 autonomy, a car that can drive itself anywhere, in any weather, without you, it’s a massive engineering and regulatory problem that’s still a long way off.
Myth 1: ADAS is the same as self-driving.
This is the single most common and dangerous myth. A lot of drivers think that because their car has adaptive cruise control or lane-keeping assist, it’s a “self-driving” car. It’s not. Advanced Driver-Assistance Systems (ADAS) are there to help you, not to replace you. These systems fall into what the Society of Automotive Engineers (SAE) International calls Level 0, 1, or 2 automation. For example, a car with adaptive cruise control (Level 1) can hold its speed and distance from the car in front, but you’re still doing all the steering and everything else. When you get a system that combines that with lane centering (Level 2), the car can handle speed and steering together, usually on the highway. But you are still the pilot. You’re expected to be watching the road and ready to take over instantly. Think of the system as a co-pilot, not the one in charge. The National Highway Traffic Safety Administration (NHTSA) [https://www.nhtsa.gov/technology-innovation/automated-vehicles-safety] is clear that for all commercially available vehicles in 2026, even with ADAS, the driver must stay engaged. Research from the AAA Foundation for Traffic Safety [https://aaafoundation.org/research/human-factors-and-road-safety/] shows over and over again that drivers are confused about what ADAS can do, which leads them to misuse it and get a false sense of security.
Myth 2: Once a car has Level 3 capabilities, I can just sleep at the wheel.
No, absolutely not. A Level 3 automated driving system is a major step up from ADAS, but it doesn’t mean you can check out completely or take a nap. The whole point of Level 3 is that the car can handle all the driving tasks, but only under very specific, limited conditions. These conditions are called the system’s Operational Design Domain (ODD). A perfect example is Mercedes-Benz’s DRIVE PILOT [https://group.mercedes-benz.com/innovation/drive-pilot-sae-level-3.html], which is approved for use on certain highways in Germany and parts of Nevada and California. It lets you take your eyes off the road, but only at speeds up to 40 mph (60 km/h in Germany), during the day, and in good weather. Here’s the catch: you have to be ready to take over within seconds if the car issues a “handover request.” This is non-negotiable. If you don’t respond, the car is designed to perform a minimal risk maneuver, like pulling itself over to a stop. The car takes over the job of watching the road, but you’re still the backup who has to be ready to jump in. It’s a big change from Level 2, where you’re always the one watching.
| Feature | Level 2 ADAS | Level 3 Automated Driving | Level 5 Full Autonomy |
|---|---|---|---|
| Driver Supervision Required | ✓ Constant attention, hands-on ready | ✓ Driver available for handover | ✗ No human intervention |
| “Eyes-off” Driving Permitted | ✗ No | ✓ Yes, under defined conditions | ✓ Yes, all conditions |
| System Monitors Environment | ✗ No (driver is primary monitor) | ✓ Yes (system monitors) | ✓ Yes |
| Operational Design Domain (ODD) | ✓ Specific conditions (e.g., highways) | ✓ Specific, limited conditions | ✓ All conceivable conditions |
| Commercial Availability in 2026 | ✓ Widespread | ✓ Limited (e.g., Mercedes-Benz in specific areas) | ✗ Not expected |
| Responsibility for Dynamic Driving Tasks | Driver | System (driver as fallback) | System |
Myth 3: All self-driving cars will be able to drive anywhere, anytime.
This myth comes from watching too many sci-fi movies and ignores the tough engineering and regulatory work that has to happen first. Even the most advanced automated systems, including those aiming for Level 4, are restricted to their Operational Design Domains (ODDs). An ODD is just the list of conditions where a system is designed to work safely. This includes everything from weather (no heavy rain), to road type (paved highways only), to location (a specific part of a city), and even time of day. So, could a Level 4 robotaxi drive you around downtown San Francisco on a sunny day? Yes. But you couldn’t expect that same system to handle a blizzard in rural Alaska. The sensors, maps, and software are all built for those specific ODDs. To get to Level 5 autonomy, where a car could truly handle any road in any condition without a human, is an incredible technical challenge. The system would need to deal with freak weather, weird debris on the road, and unpredictable people without having been pre-programmed for every single possibility. People are working on it, but widespread Level 5 is a distant dream, and definitely not happening by 2026.
Myth 4: Automated driving will eliminate all traffic accidents.
Automated driving tech should make roads a lot safer by cutting down on human mistakes, but thinking it’ll stop every single accident is just unrealistic. Yes, human error is behind most crashes, but cars can still have mechanical failures, hit a deer, or run into problems with the road itself. And these automated systems introduce their own set of problems, like software bugs, sensor failures, or cybersecurity attacks. The long transition period where we’ll have a mix of human drivers and automated cars on the road at the same time is also going to be tricky. How does an automated system deal with a driver who is texting or driving erratically? How will we learn to predict what an automated car is about to do? The RAND Corporation [https://www.rand.org/pubs/research_reports/RR2150.html] has done a lot of research on this and their work makes it clear that while we can expect huge safety gains, a world with zero accidents isn’t in the cards. We should expect a big drop in crashes, which is great, but zero isn’t going to happen.
Myth 5: Automated vehicles are just around the corner for everyone.
People are way too optimistic about when we’ll all have self-driving cars. We’re definitely seeing more Level 2 ADAS features in new cars and a few limited Level 3 systems on the market, but getting to where everyone has a Level 4 or 5 car is a huge leap. It needs better tech, new laws, public buy-in, and lower costs. Making this software safe requires millions and millions of miles of testing, both in simulators and on real roads in all sorts of conditions. Government bodies like the National Transportation Safety Board (NTSB) [https://www.ntsb.gov/safety/safety-issues/Pages/AV.aspx] are still figuring out the rules for testing, deployment, and who’s liable when something goes wrong. Public trust is also a big deal. A few high-profile accidents can scare people off and slow everything down. On top of that, the cost of the advanced sensors, powerful computers, and HD maps needed for higher automation adds a ton to a car’s sticker price right now. The tech is moving fast, but it’s smarter to expect a slow, gradual rollout of these features over many years, not some overnight revolution in driving. The timeline for widespread Level 4 robotaxis is still measured in years, and personal Level 4 or 5 cars are even further down the road.
Level 2 vs. Level 3: What’s the real difference?
The main difference is who’s responsible for watching the road. With Level 2, the driver must always supervise the system and be ready to take over instantly. With Level 3, the car’s system takes over monitoring the environment (within its specific ODD), letting the driver take their eyes off the road. But, the driver must still be available to take control when the car requests it.
Can a Level 3 car drive in bad weather?
No. Level 3 systems can only operate in their specific Operational Design Domain (ODD), which almost always means good weather. If it starts raining or snowing heavily, that’s outside the ODD, and the system will require the driver to take control.
What does “Operational Design Domain” (ODD) mean?
The ODD is simply the set of specific conditions an automated system is built to handle safely. This includes things like road types (e.g., highways only), geographic areas, speed limits, weather, and time of day. If the car goes outside its ODD, the system will disengage or require the human to take over.
Can I buy a Level 4 car today?
You can’t buy one for personal use. A few companies like Waymo and Cruise run limited Level 4 robotaxi services in geofenced areas like parts of San Francisco or Phoenix, but these aren’t for sale to the public and are strictly confined to their pre-defined ODDs.
What’s the role of government regulation in all this?
Regulations are essential to make sure these systems are deployed safely and responsibly. Government agencies are working on the standards for everything from testing and performance to cybersecurity and liability. These rules are still a work in progress as the technology itself develops.
