Asger Risom
New research models how self-driving technology could change crash outcomes over the next decade.
Autonomous vehicles may deliver meaningful public health gains even at low adoption levels, according to new research that models how self-driving technology could change crash outcomes over the next decade.
The study argues that modest penetration of automated driving could translate into tens of thousands of avoided injuries, with far larger benefits if the technology scales and performs as early data suggests.
A growing safety problem
Traffic crashes remain a major cause of injury and death in the United States. More than 120 people die on US roads each day, and in 2022 alone crashes led to over 2.6 million emergency department visits for injuries.
Beyond the human toll, road accidents impose enormous economic costs. The study cites estimates of more than $470 billion annually in medical spending and lost productivity tied to traffic fatalities and injuries.
Because most crashes stem from human error or substance use, researchers say reducing driver involvement altogether could significantly improve safety.
Modeling the impact
The analysis, published in JAMA Surgery, was conducted by a team of researchers based in Canada. They examined US road traffic injury data from 2009 to 2023 and projected trends through 2035 using linear regression.
Their approach centered on two variables: how widely autonomous vehicles are adopted, and how much safer they are than human drivers. Both factors, the authors note, remain uncertain but can be explored through scenario modeling.
Adoption scenarios
To estimate future use, the researchers applied a growth curve to model four possible adoption levels by 2035: autonomous vehicles accounting for 1%, 2.5%, 5%, or 10% of all miles driven in the US.
They paired those adoption levels with two safety assumptions. In the conservative case, autonomous vehicles were assumed to be 50% safer than human drivers. In a more optimistic scenario, based partly on early findings from companies such as Waymo, AVs were assumed to be 80% safer.
Injury reductions
Under the most cautious assumptions—1% of miles driven by AVs with a 50% safety improvement—the researchers estimated more than 67,000 injuries could be avoided between 2025 and 2035.
At the other extreme, with 10% adoption and an 80% safety advantage, the number of prevented injuries exceeded 1 million, corresponding to an overall reduction of about 3.6% in traffic-related injuries over the decade.
The authors stress that even limited adoption could yield measurable benefits, challenging the idea that self-driving cars must reach mass deployment before affecting public health.
Limits and next steps
The researchers caution that the projections rely on limited real-world safety data, as autonomous vehicles are still used at small scale. As deployment expands, the estimates will need revision.
They also suggest future studies focus more closely on highway driving, where the most severe injuries and fatalities occur, to better understand how autonomous systems perform in high-risk environments.
Sources: JAMA
