Tesla’s EV recall is bad for Its autonomous car rivals

Tesla filed a recall report with the National Highway Traffic Safety Administration for about two million vehicles. (File image)

Having virtually all of your company’s vehicles on US roads recalled over a long-running safety problem seems like it would be bad for your company. Unless you are Tesla Inc. In which case, it is more of a problem for your competitors.

Tesla just filed a recall report with the National Highway Traffic Safety Administration for about two million vehicles, its biggest ever. The regulator determined that Tesla’s driver-assistance system, dubbed Autopilot, doesn’t do enough to ensure drivers remain engaged and don’t misuse the system. Tesla plans an over-the-air software update that will install more controls and warnings.

The stock is down all of 2.6 percent as of writing this on Wednesday morning. This is, after all, Tesla. Its 75-times earnings multiple is premised, implicitly, on it eventually cracking fully autonomous driving. And such faith, tied to the genius of Chief Executive Elon Musk, is so rock solid that any apparent setback is rationalised and set aside. The horizon is what matters here, not the immediate surroundings.

This is obviously great for Tesla. Musk has been touting the imminent arrival of vehicles that drive themselves since at least 2016. They remain conspicuously absent; the company’s valuation is about $700 billion higher nonetheless.

What it isn’t great for is the wider project of autonomous driving.

Even compelling new technologies rarely, if ever, flood through society without accompanying, enabling technologies and adjustments to behaviour. That’s true for electric vehicles, where recharging requires new infrastructure and customer experience, and was true for internal combustion engine vehicles, too (see this).

Apart from enabling more leisurely rides, autonomous driving is pitched as a way to reduce accidents. What tends to get lost in there is that, even if the data could establish that AI vehicles kill fewer people than regular intelligence vehicles, the idea of being killed or injured by a robot is just more horrific. We are irrational that way, I suppose, so getting to the robotaxi future is going to require some serious handholding and a lot of baby steps.

Musk doesn’t do baby steps. His own engineers have struggled over the years to convey to him the difficulty involved in getting to true, so-called Level 5 autonomy, where driver input isn’t required at all (Autopilot is Level 2, requiring constant engagement). Regardless, Musk insists self-driving, self-financing vehicles without steering wheels are virtually here. As much as his hubris can backfire, it also underpins his successes and, importantly, the reality distortion field that made Tesla the most valuable listed automaker in the world.

It also distorts the reality in which competitors operate.

General Motors Co’s Cruise unit suffered a spectacular setback in October when California suspended its license for operating driverless vehicles in the state. State regulators accused Cruise of withholding footage of an accident in which a driverless vehicle struck a pedestrian, through no fault of its own, but then moved to the curb, dragging the unfortunate victim with it. GM is yet to say exactly what went wrong, but Cruise’s co-founder Kyle Vogt has resigned and the business faces an overhaul of its strategy and working culture.

An obvious culprit is Cruise’s aggressive expansion, including using San Francisco as its biggest testing ground; not the easiest city in which to perfect self-driving. Alex Roy, a former autonomous-driving executive and now principal at consultancy Johnson & Roy, contrasts Cruise’s dash for scale with what he calls the “quiet strength” of Waymo LLC’s self-driving product in Phoenix — an easier city to tackle — built incrementally over the course of more than a decade by the Alphabet Inc. subsidiary.

Roy sees a wider problem here of unrealistic expectations about driverless vehicles, pushing some developers to move faster than they are able. He laments a prevailing attitude that human intervention, usually by employees in remote operations centers, is regarded as a kind of failure for driverless vehicles rather than as a vital service to improve the product and encourage uptake by nervous people. Over time, the industry would of course need the number of remote operators to decrease, reflecting better technology and reducing costs. Still, as he says, “all products, especially an advanced technology product, have a customer service function.”

Cruise’s experience shows the risks of sprinting in what is a marathon. Yet all developers of autonomous driving technology operate in an environment shaped to a large degree by Musk’s repeated, grandiose promises, beginning with the name “Autopilot” itself. This shapes not just customer expectations but also financiers; Alphabet may not mind funding a measured, multi-decade strategy but most companies, and all venture capitalists, want scale, and quickly. That Tesla’s apparent setbacks, such as Wednesday’s, don’t seem to dent faith in its CEO’s vision only reinforces this corrosive dynamic.

Liam Denning is a Bloomberg Opinion columnist. Views do not represent the stand of this publication.  

Credit: Bloomberg