Potential for abuse in self-driving cars

This week’s Economist had a story about self-driving cars and their potential for abuse. Here are my thoughts on the article.

I want self-driving cars to become a reality, but given the way we handle Internet security, transparency, proper independent oversight in such a poor fashion, I’ll be uneasy about trusting self-driving car systems. I say “systems” because truly efficient self-driving cars will likely require communication and coordination with other cars, and with other decision-making systems.

The article talks about the expanded surveillance capabilities that will come self driving cars, and also mentions their potential for abuse within authoritarian states. These risks are real, especially if a car’s route information or decision-making processes can be influenced from outside the car itself. China already apparently restricts and influences the travel of dissidents on train systems by reducing their available travel credits. Surely they would want similar reach into self-driving car systems. Even if cars don’t get their marching orders from a central authority, there is potential for abuse given that so many computer systems are vulnerable to attackers.

In the US, I’d be much more comfortable if all self-driving cars had manual controls. I’d want to be able to take control in case of emergency, or hijacking. Must have manual controls and a physical kill-switch that disables the autonomous functionality. Not a software kill-switch, but something physical.