How Making Autonomous Vehicles “Less Human” Could Reduce Fatalities by Over a Third (IIHS)

How Making Autonomous Vehicles "Less Human" Could Reduce Fatalities by Over a Third (IIHS) 4

True “full self-driving vehicles” or full autonomous vehicle (AV) technology, operating at what’s known as level 4 or 5 autonomy aren’t exactly readily available at your local dealership or through an over the air software update. Progress will be arduous but the expected outcome, transformative. How much though, will the heroic effort be worth and how will it manage the issue of accident avoidance? Driver assistance systems continue to evolve, making the most of big data and artificial intelligence, with electric vehicle technology doing its part to enable the autonomous revolution.


But in a May report from the Insurance Institute for Highway Safety (IIHS), only about one-third of auto fatalities in select categories could be reduced by upcoming AVs. This is smaller number than one would expect and runs counter to the general premise that AVs will nearly eliminate road accidents. The one-third figure may be a conservative estimate but expected if self-driving vehicle systems are setup to drive the way drivers currently drive.  For better outcomes and accident avoidance, AV planning will need to factor in design philosophies consistent with “safety rather than occupant preference when they conflict” according to the report.

IIHS Twitter

Source: IIHS analysis of 5,000 crashes

A closer look at those select categories where accidents would be avoided 1/3 of the time include crashes involving sensing/perceiving (not recognizing hazards) and incapacitation (alcohol-impaired or otherwise incapacitated driver); areas where artificial intelligence could considerably improve outcomes. Another 66% of crash types would persist if designers program AVs to make errors similar to those of the human drivers of today. These include planning/deciding (poor decision-making behind traffic law adherence and defensive driving), and execution/performance (inappropriate vehicle control).

IIHS data in the report was from more than 5,000 police-reported crashes from the National Motor Vehicle Crash Causation Survey. The data was collected by the National Highway Traffic Safety Administration and is a representative sample of crashes across the U.S. in which “at least one vehicle was towed away, and emergency medical services were called to the scene.”

How Making Autonomous Vehicles "Less Human" Could Reduce Fatalities by Over a Third (IIHS) 5
IIHS Fatal Car Accidents per State

While American roadways are trending towards fewer fatalities, there remain many lives to be saved. In 2019, approximately 38,800 people lost their lives in car crashes, a 2% decline from 2018 (39,404 deaths) and a 4% decline from 2017 (40,231 deaths).  About 4.4 million people were injured seriously enough to require medical attention in crashes last year – also a 2% decrease over 2018 figures.


The IIHS report speaks to the importance of AVs being programmed to act safely in response to what they perceive. Only a third of self-driving crashes could be preventable by AVs if they are not designed to respond safely to what the perceive, which is far less than the 94% of crashes arising from driver error. Indeed driver error is the final failure in the causal chain of events that leads up to a crash.


“It’s likely that fully self-driving cars will eventually identify hazards better than people, but we found that this alone would not prevent the bulk of crashes,” according to Jessica Cicchino, IIHS vice president for research and a co-author of the report.


Then there’s the pricing for AV systems and AV-ready capabilities. How do you price systems that have the potential to drastically reduce fatalities and bodily harm while streamlining the driving experience?  Starting on July 1st, Tesla’s CEO Elon Musk says the company’s top tier “full self-driving” version of its autopilot assistance package will go up by $1,000 to $8,000. Tesla did the same in November of last fall raising the price of the system $1,000 to $7000. These changes represent added functionality including the Smart Summon feature as part of the software’s version 10 release. Musk has suggested that the self-driving feature could be an investment well worth making, with a true value of $100,000.


How Making Autonomous Vehicles "Less Human" Could Reduce Fatalities by Over a Third (IIHS) 6


Any way you price it, self-driving technology will continue to require massive infusions of research and testing dollars with the cooperation of regulators at the local and federal level. But if the technology can achieve its mission of reduced driver involvement and with greater real-world passenger safety the tangible value of such systems will be far greater than we can imagine.



Leave a Comment

Your email address will not be published. Required fields are marked *