Automation is coming for every aspect of the U.S. economy, but it gets the most attention for transport. Carmakers have for years claimed that full Level 5 automation is a short time away, and are in a competitive race to produce it through various commercial models. Mass transit is a subset of this genre with enthusiasm for automation, namely driverless trains. But the rollout of driverless buses is just as important, since they account for 47% of U.S. transit ridership.   

Automation would bring many benefits for bus service, specifically. For several years transit operators have had a labor problem: they simply can’t hire enough drivers. This December, Boston’s MBTA had so few that it couldn’t run 1 in 20 trips. St. Louis slashed frequency in response to a 130-driver shortage. TriMet, Portland’s operator, has made more severe cuts than those it made during the Great Recession. This problem also applies to school buses, with many drivers losing their jobs when schools were shuttered in 2020 and declining to reapply. 

Even when bus driver shortages aren’t occurring, labor remains the highest cost of transit operation—in high income nations, around 70% of operating expenses. There are a number of U.S. transit agencies, such as NYMTA, that face severe legacy debt due partly to high labor costs. 

These problems have occurred as autonomous vehicle (AV) technology has continued to improve, causing some agencies domestically and abroad to try autonomous busing. A series of pilots have been conducted or are still underway, generally utilizing 4-8 passenger shuttles. These include transit services in Las Vegas, Singapore, Taipei, and Providence, RI. Recently, the National Park Service announced an automated shuttle pilot to carry Yellowstone visitors. Perhaps the most interesting company is EasyMile, which created an EZ10 vehicle that’s been used in over 30 counties. 

The scaling of these buses could reduce operating costs and, according to a white paper by WSP, a transit planning consultancy, also reduce capital expenses. Automation would help buses “platoon” together, meaning travel at short distances from each other, providing similar or equivalent capacity to more expensive capital projects like light rail. 

But much like self-driving cars, long promised but with mild progress at best, fully autonomous buses are a ways off. The main hurdle is how they will interact with the street. A trial in Vienna found that buses were impeded by even ordinarily benign, undetectable obstacles. But for “closed” systems, namely bus rapid transit that is isolated from other traffic and within its own dedicated right-of-way, automation could conceivably happen sooner.

Most autonomous shuttle programs thus far have had a human present, due to the need for intervention in emergencies. In the short term, automated technology could be deployed in bus garages, according to WSP. Buses face fewer obstacles in storage areas, as opposed to uncontrolled streets where they must interact with pedestrians and other vehicles. So bus parking could prove to be a testing ground that yields innovation.

But as with trains, labor unions have opposed bus automation. The AFL-CIO’s Transportation Trades Department set forth a list of principles three years ago to forestall major changes, and included a call that “any use of automated technologies by transit agencies must be covered through the negotiation of implementing agreements with affected employees and the preservation of current collective bargaining rights,” meaning requirements that human operators always be present.

But it’s possible that automated buses, even with a human staffer on board, could prove more efficient. Bus drivers face multiple burdens besides driving, from fare collection to dealing with hostile passengers. The stress of the job undoubtedly contributes to the driver shortage, namely during the rise of other delivery jobs that don’t require so much unwanted interaction. Removing the driving requirement from on-board personnel would allow them to focus on rider-oriented tasks, while being able to assume command of the vehicle should a situation arise. In the national park example, the employee could provide customer service duties.

So even if AVs don’t totally eliminate the need for human staff in buses any time soon, they could change the nature of what staffers do on buses, to better improve rider safety and experience. But for technological and political reasons, full autonomy is far off.

This article featured additional reporting from Market Urbanism Report content staffer Ethan Finlan.