Much of the debate around automated cars centers around wrecks. Proponents say that a lot of wrecks could be prevented. Those who are against the technology say that hackers and computer errors would just cause wrecks. Which side is right?

While both arguments have merit, actual experts almost always agree that collision totals would fall. The only real debate is by how much. For instance, the IIHS claimed after one study that about a third of wrecks would be prevented by automated cars, breaking it down like this:

  1. Perception
  2. Incapacitation
  3. Prediction
  4. Planning
  5. Execution
  6. Vehicle failure and other “not human error” events

But critics of that study claim that 33% is far too low. They point out that the first two categories alone represent a third of all crashes. Planning includes things like following too close or driving too fast, and it accounts for another 40%. A computer, of course, could be programmed to never break the speed limit and always stay three seconds behind the next car. Shouldn’t that count as well?

In that light, they say that automated cars could prevent about 90% of car wrecks. With about 40,000 dying every year — not to mention vastly more people getting hurt — eliminating 90% of these events would save a massive number of lives, prevent life-changing injuries and keep costs down. Even if computer errors caused some wrecks, it’s doubtful they’d cause as many as people do.

This study sheds light on the risks we still face with human drivers. Those who get injured may need to seek compensation for their medical bills and other damages and losses.