BYD Says Fully Autonomous Cars Are Basically Impossible

CNBC had access to an interview in Mandarin – possibly a press conference – in which Li discussed the matter. To be precise, he said that "self-driving tech that's fully separated from humans is very, very far away, and basically impossible." Although the top BYD executive did not elaborate on that, he gave some tips on the issues such a technology would have to overcome.

In the same interview, Li said that it involves way more than eventually proving the technology works. He mentioned that there are other concerns related to ethics, "human psychological safety needs," and regulation. The executive seemed to be particularly concerned with who would be to blame in crashes with fatal victims. Who would be responsible for criminal and civil liabilities if the vehicles operated in autonomous mode when the incidents happened?

The truth is that solving the technology is just the tip of the iceberg in all discussions involving autonomous driving – even if, apparently, it is the most challenging element to sort out. Several surveys demonstrated most customers are still afraid of such a feature and would probably avoid using it. At the same time, overreliance on immature tech is what led to several deaths involving vehicles trying to achieve full autonomy. The only one related to a truly autonomous vehicle happened in Uber tests on March 18, 2018. All others relate to advanced driver assistance systems.

This is probably the reason Li said that BYD analyzed the logic behind pursuing full self-driving tech. Why would a car company want to offer a product that will probably put it as the target to sue in case anything goes wrong? According to the executive, BYD failed to figure it out. He was even more incisive, stating that it is a "false proposition." Ouch!

The BYD executive added that autonomous technologies would be more useful if applied to manufacturing, but it is not clear how they could help in making cars. Driving them to the parking lot after they are ready? Loading them in trucks, trains, and ships? As long as that always happened in the vehicles' Operational Design Domain (ODD), these applications would be perfectly possible. The question is how useful that would actually be. Would it pay off to keep pursuing autonomous driving for such a limited application?

BYD will either be the first company to declare it is not interested in the technology, or it will be the only one not pursuing it. Volvo, Polestar, Mercedes-Benz, Volkswagen, Ford, and several other carmakers are actually trying to achieve it – apart from Tesla. Most of them profess that their goal is to reach zero casualties with their vehicles in a given year – 2030 is the most cited. What they fail to grasp or choose to ignore is that it will be pretty tricky to beat humans as drivers.

As I wrote in a previous story, Steve DaSilva used National Highway Traffic Safety Administration (NHTSA) numbers to calculate how safe human driving was in the US. If drivers in the country rode 2,903,622,000,000 miles and crashed 5,250,837 times in 2021, that means one crash for every 552,983 miles – or no issues for 99.999819% of the way. The question the Jalopnik writer helped to pose was: can any autonomous driving tech deliver a vehicle that does not crash for at least 99.9999% of the time?

As unlikely as the numbers make it sound, several companies think it is possible. Their fear is that someone will eventually solve the puzzle, so why not try to beat them to it? So far, these companies are not doing that well. Cruise never disclosed how many miles its vehicles traveled, but it has already faced a few wrecks. It competes with Waymo to be one of the leading driverless tech companies in the world. The Google competitor is no exception, having had its vehicles involved in crashes as well. Argo AI already gave up on trying to make driverless cars possible.

Whether it is a life-saving solution or a dead-end, we're still some years from discovering. That means several companies will invest billions of dollars more in it until they can reach any conclusion. Another danger is that some may still insist that only new hardware and software solutions will make full autonomy attainable. As they are yet to be developed, that means the promise can resist for as long as companies think pursuing it is worth it. BYD already decided it isn't.