A warning to Canada: Self-driving taxis won’t be a traffic cure-all, and they ignore the real disease

Open this photo in gallery:

Peter D. Norton is associate professor in the Department of Engineering and Society at the University of Virginia. He is the author of Fighting Traffic: The Dawn of the Motor Age in the American City and Autonorama: The Illusory Promise of High-Tech Driving.

In Toronto, Mac Bauer races streetcars on foot – and he hasn’t lost yet.

Mr. Bauer isn’t interested in proving how fast he can run. His point is that Toronto’s streetcars are too slow. “I don’t like winning. I really don’t. I really, really wish these streetcars were faster than me,” he told the Guardian. “But they’re not. And this is the problem.”

Slow public transportation is not a problem for everyone. If you’re in the self-driving robotaxi business, slow streetcars and buses are a selling point for your service. Where public transportation is slow, robotaxis can move some people more quickly through the city. It’s an attractive proposition.

But robotaxis are not a cure. Instead, they are a symptom of a condition cities can’t afford to ignore: public transportation systems struggling to serve riders on streets that favour drivers. Waymo, the self-driving ride-hailing service owned by Alphabet, joined Toronto’s lobbyist registry in December, and met with senior B.C. officials in recent months. Some people want to welcome Waymos as a cure for slow traffic.

How companies are using generative AI to make self-driving vehicles safer

Editorial Board: The slow death of public transit service

But in medicine, symptom management must never be mistaken for the cure. The danger is that by relieving symptoms, a drug can excuse neglect of the underlying condition, permitting it to worsen. In city traffic, robotaxis can’t treat the underlying condition – and when cities resort to them, they let the ailment get worse.

Under current conditions, public transit in Toronto rarely offers time advantages over driving. Streetcars and buses crawl during rush hour. But for those who ride it, public transit has substantial cost advantages over hailing a cab, and over buying, fuelling, maintaining and parking a car. Robotaxis are expensive vehicles with expensive tech. They cost a lot to acquire, maintain and store, and they require remote supervision by human monitors.

If companies charged fares that reflected the full cost of the passenger’s trip, the effects of introducing robotaxis would be no worse than the effects of conventional cabs with human cab drivers, which have long served as a useful part of the mobility mix even before the arrival of the automobile. Because the cost of operating them kept fares high, they were confined to the niche roles they served best.

But to attract riders and investors, robotaxi companies charge artificially low fares. They lose money on every ride. Just how much their costs exceed their fare revenues is unclear – the companies prefer it that way – but by keeping fares low, they can divert some passengers from public transportation. In turn, the rising numbers of users attract investors who want to put their money into the companies that are most likely to grow. Any robotaxi company that charged riders the real cost of a ride would be left behind.

Artificially low fares produce artificially high demand – demand that is routinely misrepresented as a preference. But we can’t understand preference under such conditions. Sure, I would prefer to fly first class, for instance, but when I count cost into the calculation, I prefer to fly coach. Cost is part of the preference calculation, and distorting costs distorts demand.

Robotaxi companies are competing to get to No. 1, and they’re willing to lose a lot of money to get there. If a winner emerges, it will have to make good on its promises to investors and make rides pay. Given their high costs, fares will have to rise. When such a day comes, public transit services may recover some of the ridership they had lost to these artificially affordable services – but by then, such services may be too frail to rise to the occasion.

If robotaxis one day ply the streets of Toronto, those users who can pay the fares will report a better transportation experience in the city. But for public transportation passengers, the situation will be worse. Streetcars and buses already don’t get the kind of priority they would need to beat Mac Bauer in races, and as robotaxis divert some passenger travel from public transportation to passenger cars, the pressure to deny public transit the priority it requires will be all the greater. Meanwhile, the self-driving cars would contribute to the traffic congestion that slows public transit down so much.

Robotaxis need a lot of space, and the streets where they operate, and the curb space where these companies sometimes store their fleets, are expensive. Streets are administered by public agencies on behalf of the public who owns them. Such agencies have a right and a duty to exercise their authority in the public interest. At a minimum, this should mean that companies should have to pay for the road capacity they use. They pay for some of it now, through taxes and fees, but the ubiquity of empty Waymos – both moving and parked – in the streets of the cities where the company operates tells us that the public is not getting full value for the space it provides.

In Toronto, streetcars carry about 230,000 people daily. If Waymo diverted even 10 per cent of trips from streetcars to robotaxis, the consequence could be something like 23,000 more passenger cars on the city’s streets. Each car would be taken as a reason to accommodate driving to a degree that would preclude prioritizing streetcars. And each car would be another reason why Mac Bauer could count on winning his foot races with streetcars.

The question before any city is not whether to permit or forbid robotaxis; the question is what role robotaxis can best serve in the mobility mix. Robotaxis can offer a useful premium service in niche roles. But any city that sees them as a broader answer to public transportation frustrations is making an error as egregious as removing bike lanes to install curb parking. It’s an easy mistake to make: Some people eat poorly but take supplements, confident that the supplement-taking offsets the bad diet. Yet the benefit is worse than zero, because the supplement-taking deters necessary change. Robotaxis are like supplements, diverting our attention from mobility essentials.

Before robotaxis are welcomed at all, public transportation services must be so good that robotaxis will not delay them or divert passengers from them. By this standard, Toronto – or most other North American cities, for that matter – should not be welcoming large deployments of robotaxis any time soon.