A CBS Sunday Morning report promoted driverless cars as a way of improving highway safety. I’ve been saying the same thing for years, but I was surprised when the report opened with the statement, “Every year, 1.2 million people die in car accidents.” Technically, that’s true, but only 3.6 percent of them are in the United States.
According to Wikipedia (which is probably the source of the 1.2 million number), China sees 250,000 traffic fatalities per year. India is number two at more than 155,000. Despite the fact that all of Africa has only about 30 percent as many vehicles as the United States, Africa suffers 322,000 fatalities per year, with particularly high numbers in the Congo, Ethiopia, and Nigeria. Other counties with high fatalities include Brazil, Indonesia, Pakistan, Thailand, and Viet Nam.
Estimates of vehicle-kilometers traveled aren’t available for most of these countries, but fatalities per million vehicles are stunningly high in many of these countries. While the United States had just 0.15 fatalities per million vehicles in 2022, many African countries see between 20 and 70, and several Asian countries are around 4 to 9.
When compared with European countries, however, 0.15 is nothing to brag about, as Czechia, Denmark, Finland, France, Spain, Sweden, Switzerland, and the United Kingdom are all less than half of that. We do have vehicle-kilometers of travel for many European countries, and most have fatalities per billion vehicle kilometers that are about half those of the United States: the U.S. had about 8.2 fatalities per billion vehicle kilometers compared with an average of 4.4 in a dozen European countries.
Why are fatality rates so much higher in the U.S. than in Europe? No one knows exactly, but many blame the problem on the bigger cars, pickups, and SUVs that we drive. That may play a role, but other factors may be involved as well. The recent rise in U.S. pedestrian fatalities is largely due to increased homelessness. Europe may also have lower motorcycle fatalities as every European country mandates use of a helmet while many U.S. states do not. Distracted driving is a problem in both the U.S. and Europe, but may be worse in the U.S. as we don’t have good data in either area.
The good news is that U.S. fatalities and fatality rates are both declining. Last month, the National Highway Traffic Safety Administration released early estimates of 2024 fatalities, which dropped for the third straight year despite increases in driving in each of those years. Despite these drops, 2024 fatalities were still higher than in any year from 2008 to 2020.
The U.S. enjoyed a more-or-less steady decline in fatalities from 1973 (when more than 55,000 people were killed) until 2011 (when fewer than 33,000 were killed), and a decline in fatality rates from 1900 (when about 360 people died per billion vehicle miles) to 2014 (when it was less than 11). However, fatalities and rates rose until 2018, dropped slightly in 2019, then zoomed up during the pandemic, peaking at nearly 43,000 (close to 14 per billion vehicle miles) in 2021.
No one really knows why fatalities rose after 2011 (and rates rose after 2014), nor why both rose during the pandemic. Given the rise of smart phones, it seems reasonable that distracted driving was the cause of the first increase, but we can’t be sure.
One interesting thing is that the post-2011 rise was all in urban areas. Between 2009 and 2023, rural fatalities dropped by 14 percent while urban fatalities rose by 65 percent. We don’t have a breakdown of 2024 numbers by urban and rural, but urban fatalities declined in both 2022 and 2023, so it seems likely that they dropped in 2024 as well.
Self-driving cars are going to make roads safer, but it will be several more years before people can buy totally driverless cars. Until then, we need better data on what roads are most dangerous (they tend to be non-freeway arterials), what vehicles are most dangerous (it doesn’t include all SUVs), and what can be done to make them safer. We know that trying to get people to stop driving doesn’t work, but there is a lot more we can do to make driving safer.
Like everything AI, self-driving cars using AI will be the future of surface transportation. But … it will take a long time to replace the 283 million registered vehicles (average age 12.6 years) currently on US streets and highways.
If nothing else, it will give the popular press plenty to write about.
From MotorTrend
Self-Driving Cars in 2025: Where We Are Right Now According to Ride AI
Key considerations for autonomous vehicle development in 2025.
https://www.motortrend.com/news/ride-ai-2025-autonomous-driving-conference-report