Self-driving cars are closer than we/many want to accept. Closer to regulatory approval than ever too. Even with the limited (or so we think) tech currently implemented already, Tesla’s track record on autopilot (which isn’t autopilot officially) is as good as the human driving record in the USA in the terms of fatalities.

The ethical debate about hit the pavement to escape a frontal collision, or go frontal, will highly likely be settled by regulators IMHO. improving safety standards and results of crash tests will help make those decisions. A computer operating at the speed they do in a Tesla can compute whether the impact will go beyond the crumple zone or not and decide based on that. I expect that computation will become a requirement. And, of course, luckily not every time a human is also on the escape trajectory.

Programmed AI bias will be an interesting topic to follow but it will also be an awesome, and ethical, application for big data.