Tesla's Full Self-Driving system has improved dramatically, with some owners reporting thousands of miles requiring minimal driver intervention. Yet experts caution that even highly capable driver-assistance technology remains fundamentally different from true autonomous driving, and the critical question isn't whether FSD works most of the time, but what happens during the rare moments it fails.

How Good Has Tesla FSD Actually Become?

Recent real-world evidence suggests Tesla's latest Full Self-Driving (Supervised) software has reached a level of capability that would have seemed unlikely just a few years ago. One Model 3 owner completed a coast-to-coast trip exceeding 2,700 miles using FSD without a single intervention, and later accumulated more than 11,000 miles without touching the steering wheel or pedals while the system was active. A Cybertruck owner reported driving from Midland, Texas to Dallas-Fort Worth without needing to intervene at all. Independent testing by Barron's found Version 14 software dramatically smoother and more capable than earlier iterations.

Tesla has accumulated billions of miles of real-world driving data using FSD, providing the company with an extraordinary foundation for continuous improvement through over-the-air software updates. The system can now operate on city streets and highways, follow navigation routes, change lanes, negotiate intersections and roundabouts, make turns, and park. These capabilities represent a significant leap from earlier versions that struggled with basic urban driving scenarios.

Why Does the 1% Problem Matter More Than the 99%?

Despite these impressive capabilities, a critical distinction separates what FSD can do from what it actually is. Tesla formally classifies Full Self-Driving (Supervised) as a Level 2 driver-assistance system under the automation definitions used by the National Highway Traffic Safety Administration (NHTSA) and SAE International. This classification means the vehicle can control steering and acceleration or braking simultaneously, but the driver remains responsible for watching the road, obeying traffic laws, and intervening whenever necessary. You can take your hands off the wheel if the system permits it, but you cannot take your brain out of the driver's seat.

The danger lies not in FSD's typical performance but in what happens when it fails at the exact moment a driver has become convinced the vehicle no longer needs supervision. NHTSA opened an investigation in October 2025 into reports of Tesla vehicles operating with FSD proceeding through red traffic signals, entering opposing traffic lanes, or making other potentially illegal maneuvers. At the time the investigation opened, NHTSA had identified 58 incidents, including 14 crashes and 23 reported injuries. Six reports involved vehicles entering intersections against red lights and subsequently crashing, with four of those crashes resulting in injuries.

These numbers require careful context. They represent reports being investigated rather than confirmed findings that FSD caused every crash, and they must be weighed against millions of vehicles and billions of miles of FSD operation. Additionally, Tesla continuously updates the software, meaning the system someone experienced two years ago may differ substantially from what current owners use. A June 2026 crash in Katy, Texas illustrates the complexity: a Tesla Model 3 with FSD active left a 30-mph residential street and struck a house at more than 70 mph, killing someone inside. However, preliminary NTSB data showed the driver had manually overridden the system by pressing the accelerator pedal to 100 percent before the crash, demonstrating why headlines about crashes involving driver-assistance systems don't always tell the complete story.

How Should Drivers Approach Level 2 Automation Systems?

• Understand the Classification: Level 2 systems like Tesla FSD, Ford BlueCruise, and GM Super Cruise can control multiple driving functions simultaneously, but they are not autonomous vehicles. The driver remains responsible for monitoring the road and intervening at any moment.
• Recognize the Complacency Risk: The Insurance Institute for Highway Safety found that 11 of the first 14 partial-automation systems it evaluated received poor overall ratings, with a primary concern being whether these systems do enough to prevent drivers from becoming complacent while the technology handles most of the work.
• Know the Limitations: While Tesla FSD operates on diverse road types, Mercedes-Benz Drive Pilot, the only Level 3 conditional automation system certified in America, achieves eyes-off driving only under severely restricted conditions: daytime, clear weather, no construction, heavy traffic, and speeds below 40 mph on approved highways.

Ford's BlueCruise and GM's Super Cruise take different approaches to Level 2 automation. Both allow hands-free driving under approved conditions while using cameras to monitor whether the driver is watching the road. GM limits Super Cruise primarily to compatible mapped roads and uses cameras, radar, GPS, and precision mapping, with customers having driven more than one billion miles using the system. Ford's BlueCruise monitors the driver's eye gaze and head position and warns if attention wanders. However, NHTSA escalated an investigation into Ford BlueCruise-equipped Mustang Mach-Es after crashes involving stationary vehicles in the roadway, including fatal crashes.

The broader industry concern extends beyond Tesla. The Insurance Institute for Highway Safety raised concerns about driver-assistance systems across manufacturers, including Tesla, Ford, GM, and Mercedes-Benz, finding shortcomings of various kinds. The fundamental problem these safety organizations are trying to address is not simply whether these systems can steer a car successfully, but whether they adequately prevent humans from becoming complacent while the technology performs most of the driving work.

True autonomous driving exists in a different category entirely. Waymo operates Level 4 autonomous ride-hailing vehicles without a human driver in specific geographic areas. Through March 2026, Waymo reported more than 220 million miles with no human driver aboard, with published data showing substantially fewer injury and serious-injury crashes than human-driver benchmarks in the cities where it operates. Importantly, Waymo achieves this not by claiming its system can handle virtually any road, but by severely limiting when and where it assumes responsibility, operating only within carefully defined areas and conditions.

Tesla has chosen perhaps the most ambitious route: building a system capable of handling an enormous variety of roads and situations using camera-based technology. This approach generates both impressive real-world results and legitimate safety questions. The distinction between a remarkably capable driver-assistance system and true autonomy remains crucial for every driver to understand, especially as these technologies continue to improve and become more prevalent on American roads.