Driverless vehicles will take at least ten years to deploy over large areas, and adoption will occur at different rates across the country, according to an MIT research brief.
“We expect that fully automated driving will be restricted to limited geographic regions and climates for at least the next decade,” the authors write.
The brief, released Wednesday, was published by a university task force created two years ago to examine the future of work during an “age of innovation.”
As such, a significant portion of the report is dedicated to the jobs impact of autonomous deployment, an impact the authors state will not be as dire as some have predicted.
“The AV transition will not be jobless,” the report states, and new job opportunities will arise, such as remote operation of vehicles.
The quality of these jobs is uncertain, the authors acknowledge, and depends in part on public policies aimed at retraining workers, especially in trucking.
Hype vs. reality
Ever since Google launched its self-driving car project in 2011, predictions about autonomous vehicle deployment have toggled between champions who make ambitious claims about the imminent deployment of autonomous trucks and cars, and skeptics calling for a reality check.
Widely viewed as the easiest sector to automate, the self-driving trucking market is nevertheless caught between these competing visions.
In the wake of the coronavirus pandemic and moderate investor enthusiasm for technology investment, the industry today is undergoing a wave of acquisitions and layoffs.
At the same time, many companies are pushing forward with new industry partnerships and test operations, touting an optimistic vision of a soon-to-be-autonomous future.
The autonomous trucking outfit TuSimple stated earlier this month that it aims to establish an autonomous trucking service from Los Angeles to Jacksonville, Florida, by 2023, followed by a nationwide freight network in 2024. And autonomous trucking startup Aurora Innovation on Monday showed an autonomous driving system installed in a Peterbilt 579 model that it plans to test in Texas.
Cost, technology development hurdles to market penetration
The MIT authors, whose conclusions draw on existing research, remain skeptical about rapid near-term adoption. Although many tech startups cite regulation as the primary hurdle to autonomous vehicle deployment, the authors point to technical glitches and cost as challenges the self-driving car, transit and truck sectors have yet to overcome.
“In terms of technical knowledge, the expansion of automated vehicle systems is likely to be quite slow, because there is no guarantee that improvements in driving performance will happen linearly or predictably in these varying applications.”
The cost of hardware, sensors and vehicle systems, as well as infrastructure, are often underestimated, the authors state, with research showing the market uptake of AV technology sometimes underestimates current costs for high levels of automation “by almost an order of magnitude.”
For these reasons, the authors estimate a “a slow shift” toward driverless systems with “only limited use” even in the trucking space expected by 2030.
On the subject of job losses in freight, the authors say that humans “won’t disappear” from truck fleets, but they will change roles to incorporate supervision of automation as part of the job.
Workforce education and training must be part of “a holistic response to technology and mobility change,” the report states. “Employment shifts are most likely to affect trucking jobs first, and taxi and transit jobs at a later date.”
Related stories:
TuSimple revenue falls short of expectations
Fear not the robo-apocalpyse, analysis of autonomous trucking suggests
Another autonomous trucking startup pushes back against Andrew Yang
