Robotaxi revolution has one significant flaw: Cost

A billion dollars. That's how much Uber has spent on developing driverless technology so far. The hefty figure is hardly surprising, given most of the company's revenue cash goes toward paying human drivers. Road-going robots promise to fix this, ultimately boosting Uber's balance sheets.

The ride-hailing giant isn't alone. Some of the world's largest automakers and tech manufacturers are betting on a future where self-driving technology is profitable and ubiquitous.

Why shouldn't it be? Without drivers to pay, businesses save. Some of those savings are passed on to consumers through lower fares that undercut the cost of traditional car ownership. Society subsequently abandons said ownership in droves, opting instead for robotaxis. At least that's what we're told. The reality, however, is very different.

MIT RESEARCH REPORT: Autonomous Vehicles and Public Health: High Cost or High Opportunity Cost?

For one thing, driverless does not mean humanless. Machines trim the need for human labor but seldom, if ever, purge that need entirely. The reason is simple. Algorithms err, and when they do, the results can — particularly in industries such as transportation — be deadly. This seldom-discussed reality explains why regulators have never signed off on safety-critical systems being used without human oversight. Public servants know that for all their virtues, machines can't be trusted to always get it right all the time.

Driverless car developers know this, too. Their solution? Remote control. The idea is simple. When a robotaxi malfunctions, a human intervenes to ensure passenger safety. However, rather than sitting inside the car, these "teleoperators" are located in a command center far away. Think of it as air traffic control for cars. Yet, while the setup may be lauded on safety grounds, it raises an interesting question — how many robotaxis should one teleoperator watch? The ideal answer for bean counters is as many as possible. But can one person really help a distressed vehicle while keeping an eye on several others? Would you feel safe knowing your teleoperator is watching 10, 20 or perhaps 100 other robotaxis?

Yet the real problem posed by driverless technology isn't related to its algorithmic imperfections or safety oversight concerns. Rather, the price. Even if the technology were flawless, the cost of hailing a robotaxi would still be pricier than personal vehicle ownership today.

The reason? The taxi industry is inefficient — so inefficient that cabbies spend only a fraction of their time earning fares. The rest is spent finding them. In cities such as Beijing, 40 percent of taxi miles are driven with the back seat vacant. In Seattle, that figure rises to 60 percent. This improper matching of supply with demand carries consequences. For human cabbies today, it means less take-home pay; for driverless cab companies tomorrow, less savings that can be passed on to consumers. And less savings ultimately impacts consumer willingness to abandon traditional car ownership.

This reality seems lost on the likes of Uber. Company execs seem to spend much of their time rehashing corporate talking points about the benefits of self-driving technology. In a driverless world, emissions will drop, congestion will ease, and productivity will rise.

About 1.3 million people die every year in road accidents. Countless more are injured. Most of the blame lies with human drivers for doing things such as texting while driving, drinking and driving, and napping while driving. Driverless technology promises to fix this. Machines, after all, don't get distracted, drunk or drowsy. The result? A solution to what the World Health Organization has called a public health crisis.

Driverless technology may well deliver on all these promises. Yet realizing this reality demands cost competitiveness with today's car ownership. That's unlikely if current market inefficiencies persist. As long as they do, don't expect investment in Uber— or any self-driving car company for that matter — to pay off.