The California Assembly passed AB 316 by a vote of 54-3. Included in the bill is a provision requiring a human driver be present in autonomous trucks while testing, which was lobbied for by the Teamsters’ union. This is only an Assembly Bill: In order to become law it would need a Senate equivalent and the signature of the governor.

Previously, the San Francisco supervisors, also under lobbying from the Teamsters, refused a petition from Waymo to convert warehouse space to parking spaces for employees at a new facility in an industrial section of SW San Francisco. The parking had already been overwhelming approved by lower committees, but descriptions of the meeting suggest that the conflict between the robotaxi companies and the city has escalated, and the rejection of this parking application gives the appearance of being about that, and not whether these parking spaces are appropriate.

Robocar developers have always anticipated that there would be backlash as the technology became more real. Some of the backlash will be emotional, or from those who feel commercially threatened by the technology. Some will be legitimate, regarding the public safety and road disruption issues which are expected with the piloting of emerging technology like this. The issue of job preservation — many Teamsters drive for a living — has people on both sides, with sympathy for those who feel a threat to their career, but concerns of efforts to protect a dangerous and unpleasant job simply so jobs can be maintained. Such efforts have not usually withstood the path of history, nor has it been viewed as good in hindsight when they did.

For the trucking, it’s obviously ridiculous, long-term, to have drivers in trucks driven by computers. All teams currently do this in order to maintain safety. Their systems are not mature and they know it. But all the plans depend on removing that driver at some point, as has already been done for robotaxis in the city. The law, if passed, requires that driver be there until at least 2029, when the DMV can submit a report with recommendations about removing the driver recommendation.

Outside the state, several trucking companies are already doing tests with no driver, and one company, Gatik, does regular deliveries every day with no safety driver in the vehicle. Waiting until 2029 in California essentially pushes the technology out of the state, even though many of the companies have substantial presence there. Once something is banned, it becomes very difficult to “un-ban” because the officials who remove the ban effectively take responsibility if something goes wrong. As such, they are afraid to do it. They don’t want to take risks, even justified risks, and pay the consequences.

Indeed, the companies developing trucks have all protested this proposed law, and hope the governor declines to sign the bill. California is the world leader in self-driving technology, and has to think seriously about deliberately discarding that lead because of a union seeking to preserve jobs for its members. They should investigate other ways to ease the burden on those union members. The harsh truth is that 2,000 people die every year in truck crashes in the USA — many of them truck drivers. Preserving jobs is one thing, but preserving a job that kills so many is harder to justify. It’s different from preserving jobs which cause no deaths.

SF vs. Robotaxi

It is likely that the battle in San Francisco will continue to escalate. The state has the right to regulate the roads, not the city. This has frustrated the city which wants more say in how the robotaxis operate there. The city has been frustrated by its powerlessness, and it seems they feel that the robotaxi companies have exploited it and not paid enough respect to the city. Whether that’s true or not, feelings and perceptions matter more in this sort of fight, and the city is not at all without tools to get its way with companies operating there.

The jurisdictional line between the state and city is there for a reason. If each city had too much power for different rules of the road, it could create chaos for drivers and companies with a difficult patchwork of regulators to deal with. Waymo and Cruise are based in the SF Bay Area and that city is by far the best testing ground for both of them, but they may be starting to regret the issues that come with it. San Francisco, in spite of being the big city in the high-tech capital of the world, has a surprisingly luddite history with being the testing ground for the new technologies made there. That is its choice, but it makes it harder for companies to work there.

In the past, that’s been no problem for SF. It had more than enough cachet to scare away a few firms. Today, the move to work-from-home has emptied the SF downtown more than any city in the USA, and it needs to be wary about making companies more afraid to work there. SF has shown that even if it can’t regulate your technology, it will use its more mundane powers, like zoning approval, to get what it wants. And they should win this fight if they want to, and push out the companies — the real question is why they should want to. Yes, being the beta tester of technology like this comes with issues. But a study by SF’s transit agency where they asked their drivers to log all issues with the robotaxis found surprisingly few events. The companies are goofing up, and having many teething problems, but it seems to be anecdotes, not a pattern, but it’s being confused for a pattern.

The city needs to decide just how many teething problems it can tolerate, and then grit their teeth and tolerate them. As long as people are not getting hurt, the technology holds the promise of reducing road risk greatly in the future, and a few snafus are arguably well worth it, for society, and even for the city.

A passive-aggressive battle where cities are eager to find any other method to get rid of these companies isn’t good for either party. The sides should reconcile, or split, costly as that might be. Waymo may already be hoping for a better time in Los Angeles — that remains to be seen. Arizona and Texas have already shown that they are ready to move forward. The problem is that San Francisco didn’t ask for the tech to come to the city — it was born there so it was never a welcome immigrant. A background battle won’t hurt anybody.

Waymo will resolve its parking problem. Indeed, one obvious solution is to use a staging lot and have Waymo vehicles shuttle employees from it. It’s a bit inconvenient, but a good test of a mode that will be necessary when robotaxis get ready to handle things like stadiums and large buildings with high peak traffic flows. Doing it at small scale is more expensive — you don’t want employees to have to wait more than a few minutes for a shuttle, but this is the time when these costs can be handled. But finding a way to get SF on board is important in the long run.