Auto / Tech

New study shows self-driving cars may only stop one-third of crashes

A new study from the Insurance Institute for Highway Safety (IIHS) suggests they’ll only be able to stop one-third of car crashes.

Technology can solve a lot of problems in the world today, from workplace inefficiencies to everyday hazards. Sometimes, though, people may put a little too much faith in the potential of some technologies. Although some advances can help quite a bit, they can’t always help as much as we thought.

Self-driving cars are one of the most talked-about technologies today. A lot of conversations about them revolves around how they could improve road safety, but they may not make that much of a difference. A new study from the Insurance Institute for Highway Safety (IIHS) suggests they’ll only be able to stop one-third of car crashes.

The Safety Potential of Self-Driving Cars

The idea that self-driving cars will make roads safer isn’t an unfounded one at all. Traffic accidents often happen because of poor driving habits, which a machine can’t have by nature. When you consider how much human error plays into crashes, driverless vehicles look like an attractive option.

As much as 94% of car crashes are the result of human error like speeding or drunk driving. It stands to reason, then, that removing the human element from driving would reduce around 94% of accidents. An AI system can’t get drunk or distracted, so it’s a safer driver by design.

Machines often have faster reflexes than humans do, too, and quick reactions are essential for driving. If a hazard appears out of nowhere, a person may not be able to react in time to avoid it, but a robot could. You can reduce almost everything down to data, and AI can process data far faster than a human.

Why Can’t Self-Driving Cars Prevent All Crashes?

If human error is to blame for 94% of crashes, then why can driverless cars only stop 34% of them? According to the IIHS study, the AI systems in these vehicles can still make some of the same errors. Better perception and the inability to be distracted aren’t the only factors at play.

The study determined that self-driving cars can prevent all accidents that result from poor perception or distraction. Seeing as distracted driving kills eight people a day, that’s a significant step forward. That still doesn’t account for all traffic collisions, though.

While driverless cars could always obey the law, that doesn’t mean they would be perfect drivers. Even an advanced AI may not be able to predict behavior like jaywalking or sudden lane shifts. People act irrationally, and irrational actions would be challenging for an AI to predict.

What these systems prioritize could also lead to accidents. If they pursue passenger preferences and convenience over safety, they could make riskier choices. To avoid these kinds of situations, manufacturers would have to program driverless cars always to prioritize safety.

Delphi-Self-Driving Cars
Delphi is one of many working on self-driving cars

Dangers of Driverless Technologies

These flaws in self-driving cars aren’t unavoidable. As the technology progresses, they will become safer and be able to address these issues. Getting to that point, though, may be a bumpy road.

Most manufacturers are slowly adding more self-driving technologies to their vehicles, creeping towards full autonomy. These features, like lane assist and automatic braking, can be helpful but may have some unintended consequences. It seems strange, but if drivers feel too safe, it could lead to riskier driving.

When people have all these features at their disposal, they could become complacent. They may not pay as much attention to the road, putting too much faith in these driver-assist technologies. As a result, partial autonomy in cars could increase distracted driving, as counterintuitive as it seems.

There’s also the issue of other human drivers endangering self-driving cars. Even if a driverless vehicle is perfectly safe, other people on the road can still make mistakes that affect it. Self-driving cars won’t be as safe as they can until they make up all traffic.

Building Safer Self-Driving Cars

Manufacturers and researchers will have to make some adjustments to driverless cars before they’re road-ready. This could mean making some sacrifices in profitability by preferring safety over convenience. Cars that always follow the law and are patient would be safer, but customers may not want that.

Think about how you drive when you’re running late to an event. Since you’d have to make up for lost time, you’d probably break the speed limit and drive more aggressively. A safe self-driving car wouldn’t allow you to do that, so you may not want to buy one.

Even preventing the 34% of crashes that the IIHS study noted will take some improvement. For driverless cars to avoid all of those, they would need near-perfect sensors that would never malfunction. Given the current state of object perception in self-driving vehicles, getting to that point could take some time.

Other Factors in Addressing Road Hazards

Creating better self-driving cars isn’t the only necessary step in moving towards better traffic safety. Upgrading other parts of the road would help reduce risks and get to a safer future sooner. Everything from stoplights to the infrastructure on the roadside could use the IoT treatment.

Adding IoT functionality to these objects would create a more connected traffic system. Stoplights could adjust according to real-time traffic data, and edge computing could help driverless cars navigate. This interconnected system would lead to a much safer driving environment.

One of the most substantial challenges facing self-driving cars is environmental factors like dangerous weather. A faster weather-reporting system could help by predicting and alerting connected road systems. Everything from cars to lights could then adjust to prepare for inclement weather.

Driving Towards Safer Roads

Building a safe road network isn’t an impossible goal, but it is a challenging one. Self-driving cars have a lot of potential, but people shouldn’t raise their expectations too high. As promising as some technology is, it can only get us so far.

There’s still a lot of research and development needed before self-driving cars reach their full potential. Even after that point, they won’t be entirely safe until all vehicles on the road are autonomous. Until then, drivers still need to pay attention while driving in all circumstances.

What do you think of self-driving cars and their technology? Let us know in the comments below or on Twitter, or Facebook. You can also comment on our MeWe page by joining the MeWe social network.

Last Updated on

To Top