Self-Driving Cars and Why They Should Scare You

By David Gewirtz

The problem with self-driving cars, not to overstate the obvious, is that they’re self-driving. There’s an AI (artificial intelligence) inside, taking the place of a human driver. That AI receives sensor data from the road around the vehicle, and also interacts with data from the Internet, like maps, weather, and traffic conditions.

Because the vehicle is moving, it can’t expect to rely on a constant, uninterrupted stream of real-time Internet data. When the vehicle goes under a bridge, into a tunnel, or travels one of those big open stretches between gas stations, Internet access may be spotty. The AI has to be self-contained enough to make dynamic, real-time decisions, with only intermittent support from the cloud.

The array of sensors placed around the vehicle are the eyes and ears of the AI. By combining and analyzing data acquired in real-time from optical cameras, radar, ultrasonic, and lidar (Light Detection and Ranging) sensors, the vehicle’s intelligence can develop a situational awareness, which autonomous vehicle designers say will keep the vehicles safe on the road.

There are substantial benefits to autonomous driving technology. For the elderly, or merely just the tired returning home after a long day, autonomous AI-based vehicles may prove safer than human drivers. If a blood/alcohol level indicating impairment is sensed, future cars will take the wheel and drive (or pull over) rather than letting their owners drive drunk. With ...