June 29, 2022

Is safety taking a back seat to automation?


There’s no question that the nearing prospect of autonomous vehicles is exciting.

If we cast our eye across any car feature or news article on the upcoming autonomous vehicle revolution, it’s hard not to get tantalized by the fact that a technology freeing people from the driving duty is almost within our grasp.

Automakers are not only in competition to deliver automated driving features, but some are also visibly aggressive in labelling their brands and products with higher levels of autonomy. Today, “autonomous driving” and “Level 4/Level5” make the best headlines, and I don’t blame those companies for trying to keep up with the trend.

However, consumer excitement sometimes turns to frustration when companies overpromise on autonomy and underdeliver on safety. A recent example is Elon Musk’s Crash Course, a documentary from the New York Times looking at Tesla’s approach to communicating the realities of vehicle autonomy and the real-world implications of those decisions.

The documentary taps into a growing conversation online around #autonowashing, the practice of selling vehicles on the promise of full self-driving (FSD) when in fact all they are capable of is assisting the driver on a limited basis, often described as advanced driver assistance systems (ADAS).

To me, packaging ADAS into FSD is just one part of the issue. Surely, it’s irresponsible to sell something called “Autopilot” without making sure the consumers understand that they still need to pay a substantial amount of attention to the vehicle and its surroundings. But there is something else that’s more concerning.

What’s the logic behind selling cars knowing that their automated driving functionalities will only work in “most” cases? It’s placing autonomy before safety. Whether a vehicle claims to have FSD or ADAS doesn’t really matter here. What matters is that the automated features it offers are likely far from being error-proof. The third round of testing by American Automobile Association (AAA) found that vehicles with an active ADAS still failed to consistently avoid crashes.

Of course, no system can be 100% error proof. The elephant in the room is carmakers understand that there are ways to reduce errors, such as diversifying the sensor suite and adopting newer technologies, but some of them have focused on productizing features without designing and building them to achieve as few errors as possible.

As the CEO of a lidar company, I like to tell people that lidar is very useful even if it raises the vehicle safety level from 99% to just 99.9% or 99.99%. A 0.9% or 0.99% difference might seem negligible, but if you think about reducing risk from 1 out of 100 times to 1 out of 1,000 or 10,000, it’s a huge leap!

With the mass-market deployment of lidar just starting to roll out, we have yet to collect evidence how much safer lidar-enabled vehicles will be. What’s certain is that the consumer appetite for higher levels of safety may never be satisfied.

According to AAA’s latest study, only a small fraction of people agreed that manufacturers should focus on developing self-driving vehicles. The majority of drivers are more interested in seeing current vehicle safety systems improve.

Instead of saying safety needs to be placed before autonomy, I think safety is at the core of autonomy. If an automated system is not safer than a human driver, then its benefits will be very limited. How can we tell a driver to sit back and relax while they might need to take over any moment?

By not placing enough emphasis on safety, we’re at risk of missing out on helping motorists build trust in ADAS and FSD, which could hurt the entire industry.

According to a PAVE poll in 2020, those who have experience with ADAS tend to show a more positive sentiment towards autonomous driving. However, according to AAA, public acceptance of ADAS is still quite low. Half of the survey population still preferred no or low automation levels (level 0 – 2) in their vehicles over the next couple of years, even if there is little cost difference.

Therefore, improving the ADAS experience for today’s motorists by giving them a stronger sense of safety is the key to preparing our society for a future where FSD is mature enough for consumer use.

I’d like to thank the automakers that do focus on building their reputations on safety. I believe with the customer loyalty they have accumulated, they will be the ones winning the race to getting everyone into a smarter car without safety concerns.