Level 2

Level 2

Level 2 refers to a car that can control steering and speed simultaneously by itself, while the human at the wheel still has to pay attention at all times. It is the second of six stages on an internationally used scale for automated driving.

How much a car takes over by itself is described on a scale from 0 to 5. This scale comes from the American engineering association SAE and is used worldwide. Level 2 is the stage at which almost all of today’s assistance systems operate. The vehicle keeps itself in the lane and simultaneously regulates speed, for example the distance to the car ahead. The human at the wheel, however, remains responsible the entire time and must observe traffic. They may briefly loosen their grip, but must not give up their attention.

The stage where the human is liable

The most important difference between Level 2 and Level 3 is not the technology, but the responsibility. Up to Level 2, the driver is always legally at fault if something happens. From Level 3 onward, they may look away under specified conditions, and the manufacturer bears responsibility during that time. That is precisely why carmakers stick with Level 2 for as long as possible.

This also explains a common misunderstanding. A system can be technically very good and still only be called Level 2. Tesla's “Full Self-Driving” is a well-known example: the name sounds like full autonomy, but the classification is Level 2. The driver must be able to intervene at any time.

This stage is risky precisely because of this in-between position. If a car drives flawlessly for minutes on end, the human’s attention almost inevitably drops. Experts call this overtrust. Studies on accidents involving assistance systems almost always revolve around this point.

Cameras, radar, and a look at the driver

A Level 2 system consists of three parts: sensors, software, and actuators. The sensors are usually cameras behind the windshield, along with radar sensors and sometimes ultrasound. They provide a continuous picture of the surroundings: lane markings, other vehicles, distances, speeds.

The software evaluates this data and is often a neural network, that is, a program trained on many examples. It recognizes, for example, where the lane runs, even if the lines are faded. From this assessment, the system calculates how strongly to steer or brake. The actuators carry this out: the electric steering motor and the brake.

Almost always, driver monitoring is added on top. Some systems only measure whether a hand is exerting pressure on the steering wheel. Newer ones use an interior camera and check whether the eyes are directed at the road. If the driver does not react, the car issues a warning and switches off the function after a few seconds.

From production car to quarterly figures

In everyday life, one encounters Level 2 under brand names. VW calls it Travel Assist, Mercedes Distronic with steering assistant, BMW Driving Assistant Professional, Tesla Autopilot. Behind each of these lies the same basic idea: lane keeping plus distance control. Such systems are no longer luxury equipment and can be ordered in many compact cars.

In business news, the term comes up in connection with suppliers and chips. Companies like Mobileye, Nvidia, or Bosch sell the computing platforms for such systems. Analysts therefore estimate how many new cars will have Level 2 on board in the future. For manufacturers, these systems are attractive because they can be sold as an add-on cost or as a subscription.

Another term that appears in reports is “Level 2+”. This is not an official part of the SAE scale, but marketing. It refers to systems that can do more than pure lane keeping, such as overtaking on their own. Legally, they remain Level 2, and responsibility stays with the human.

Subscribe free. Unsubscribe the second it sucks.

High-signal news across AI, business, UX, and tech. Every morning.