Fail-operational

Fail-operational is a design approach where, even if a single part or component fails, the system doesn't stop completely but keeps operating at a safe level.

An airplane is the easiest way to picture this. A twin-engine airliner doesn't crash just because one engine fails. It keeps flying on the remaining engine and lands safely at the nearest airport. That's fail-operational. By contrast, there's a different concept for systems that simply stop working when something goes wrong—like an elevator that halts in place during a power outage—to prevent accidents. That's a 'safer stop,' not 'continuing to operate.'

Companies building cars without steering wheels pay especially close attention to this concept. In an ordinary car, if there's a problem with the brakes or steering, a human driver can grab the wheel and pull over to the shoulder. But in a car with no steering wheel at all, there's no person to take over in that moment. That's why brakes, steering, power, and sensors—the critical systems—need to be built in multiple redundant layers, so that if one fails, another can immediately take over and keep the car moving safely to a safe location.

Asking a chatbot the following question can make the concept click:

"Explain the difference between fail-safe and fail-operational in automotive design, using a robotaxi with no steering wheel as an example."

Whenever you get the two concepts mixed up, it's worth revisiting this prompt to check again.