The Risks of Self-Driving Cars

Why it’s time to hit the brakes on automated vehicles

By Kohava Mendelsohn

What if we never had to worry about driving again? We wouldn’t have to focus on the road during our commutes, gaining more valuable time. There would be no more accidents due to human error. We could even help with climate change by making every vehicle only use exactly the amount of energy it requires. These are the promises of self-driving cars.

But what are the drawbacks? What are the risks? Will self-driving cars actually deliver on these incredible promises? I’m not convinced.

First, what do we mean when we say “self-driving”? According to the Society of Automotive Engineers, there are actually 5 levels of automation in cars. Level 5 cars are fully automated and “can drive everywhere in all conditions”. Some self-driving car companies are only promising level 3 (the human must take over when warned by the

car) or level 4 (full automation only in specific conditions) cars. Level 5 cars are still a ways away, with estimates ranging from 2030 to 2050.

But in 2015, just five years ago, there were promises from Toyota, Google, and even BMW that there would be highly automated vehicles on the road by 2020. And, while there are certainly some self-driving cars on the road here in 2023, they are neither as advanced nor as numerous as the technological landslide that was promised. Companies make big claims to get future hype and investments, then fail to deliver on them down the road. They have proven we cannot necessarily trust their promises.

But surely self-driving cars will be safer? In 2015, the United States National Highway Traffic Safety Administration stated that 94% of all crashes were a result of human error. Self-driving car companies promise to fix this.

However, much of the Artificial Intelligence that is controlling self-driving cars are “black boxes”; we don’t know what’s going on inside them. Engineers feed them tons of data and they behave in a way that matches predictions lent to by the data they are fed. If they see something in the real world that they haven’t seen before, they will have no idea how to respond. As David Zipper, a Visiting Fellow at the Harvard Kennedy School studying road safety and transportation data, puts it, “people are actually quite talented at being able to handle unexpected things, which computer systems are often not.”

We have no guarantees that self-driving cars will make our roads safer. Up until this point, we have seen Tesla have multiple recalls, crashes, and misleading claims. Additionally, automated vehicle companies say their cars are safer, however they aren’t testing them in conditions similar to where most car accidents actually happen. As the New York Times says, these companies “say they are making driving safer, but verifying these claims is difficult.”

What about climate change? It’s not looking so good on that front either. Easier access to driving will allow more people to use cars who before had to take public transit. Additionally, when less mental energy is required to drive, people will drive places more, as demonstrated in a 2018 study where participants drove 83% more miles when given private chauffeurs.

More cars travelling further will not help the environment, even if these are electric vehicles. Electric cars still create huge amounts of emissions in their manufacturing process, including vast mining operations to procure materials for their huge batteries. The electricity that would power those cars in much of the world is also still from nonrenewable sources. A truly sustainable future prioritises public transit options for everyone instead of multitudes of individual cars.

We have already spent over $200 billion dollars on self-driving cars. Autonomous vehicle companies continue to overpromise and underdeliver, and their resultant products aren’t necessarily that beneficial. It’s time we ask ourselves, are self-driving cars the solution to a real problem, or another over-hyped, unsafe, harmful toy for the wealthy?