Self-Driving Cars and the Reduction of Fatalities in the Near Future

TL;DR: Most of the reduction in vehicle fatalities will only come once self-driving cars extend beyond cities. Regardless, they will transform major cities in other ways, including reduced congestion, increased productivity, new business ideas, and so on.

As self-driving cars hit the streets in the coming years, it is likely that they will start out in major cities. This is because most (all?) implementations currently require accurate 3D maps before going out on their own. Predictably, large cities would be primary targets because of the high density of customers relative to the mapping area.

In discussing the transformation of self-driving cars, many people—myself included—have mentioned how many lives could be saved once they start rolling out. Usually, the number is around 35,000 lives per year in the US. However, we often leave out a key part in the analysis: if, at first, self-driving cars are only limited to urban areas, how many lives will actually be saved?

To answer this, I gathered 2015 data from the Census website as well as the NHTSA Fatality Analysis Reporting System. The figure below shows the number of vehicle fatalities in 2015 for US cities with 1 million or more people. In total, these cities include 1,571 fatalities, representing less than 5% of the total vehicle fatalities in the US.

Number of vehicle fatalities in 2015 for US cities with 1 million or more people

If we extend this to US cities with 500 thousand or more people, it still only represents about 9% of US vehicle fatalities.

Number of vehicle fatalities in 2015 for US cities with 500 thousand or more people. Note: You can find more detailed charts here.

I think the most revealing comparison is between the number of rural and urban fatalities: slightly less than half of fatalities occur in what the NHTSA considers urban areas.

Number of US vehicle fatalities in 2015 by land use

To be clear, I am very excited about self-driving cars and how they will positively transform our world. Still, it’s important to paint an accurate picture when creating expectations for the near future. Unfortunately, most of the reduction in vehicle fatalities will only come once self-driving cars extend beyond major cities or urban areas in general. Regardless, self-driving cars will revolutionize our cities in other ways, such as reduced congestion, increased productivity, new business ideas, and so on.

Please share your thoughts in the comments below!