How Uber suffered the deadly consequences of poor communication
Uber’s communication breakdown, views on driverless down under and a pretty “suite” new travel concept: read our weekly analysis of the most important news in automated driving!
Not every news week is “one giant step for driverless mobility”. But if you dig deep enough, you’ll always find an absolute gem of a story. And digging deep is exactly what Business Insider (BI) has been doing. In an exclusive last week it was able to shed more light on Uber’s fatal crash in March of this year. It’s subscription only but thankfully ARS technica has reported the gist of it.
It’s been common knowledge for a while that there was something flawed about the way Uber’s autonomous driving software was tuned at the time of the crash. Well, Business Insider’s Julie Bort eventually got to the bottom of why this was the case: the self-driving team was getting ready to demo for the newly appointed CEO Dara Khosrowshahi.
Eager to give the boss as smooth a ride as possible, engineers were encouraged to think about "rider experience metrics". Since “false” emergency stops can really put a dampener on your experience, the company decided to disable the car’s emergency braking maneuvers and rely instead on safety drivers intervening in an imminent collision. The problem was that the system wasn’t designed to alert the driver of this: a communication breakdown that ultimately culminated in a tragic fatality.
According to the BI report, poor communication is rife throughout Uber’s ranks – even within the same self-driving unit. One engineer told BI that "one group doesn't know what the other is doing." The report also alludes to infighting between the hardware team in Pittsburg (“a bunch of academics”) and the software team in San Francisco ("whiny and ungrateful").
Now Uber has dealt with a few calamities when it comes to PR concerning the inner workings of the company. But I, like many, always felt that there is a buffer between questionable company practices and the safety of the technology itself. I was wrong. This report proves that company culture can have a direct effect on public safety. Lest we ever forget the crucial role humans play in driverless tech.
Surveying the driverless landscape
It’s difficult to say how events like that fatality affects public acceptance. But occasionally we can take a temperature reading via public surveys. Last week saw two such survey results come in.
First, to the UK, which has just got the go-ahead for a pioneering multi-million pound research center for the development and testing of driverless cars. That said, according to Fujitsu’s Technology in a Transforming Britain report, 41% of the UK public would feel uncomfortable being picked up by one. And only 18% would put their kids in one.
Meanwhile in the former territory down under, the results of the Annual Victorian Self-Driving Car Survey came in. 18,000 Australians have spoken and in a nutshell, this is what they had to say.
“Fewer motorists (than last year) want a fully self-driving car, while more motorists want the latest driver assistance features,” the report concludes. Also, “respondents were more likely to accept driverless capability in their next car if it was restricted to freeway driving.” This acceptance grew again if the driver was required to monitor the situation at all times.
I actually interpret this as pretty promising. Remember, trust is not all about full driverless capability – Level 5 is still a long way off and for many, still somewhat intangible. There are important steps in between: advanced driver assistance systems and scenario specific use cases. As the Australians have shown, being skeptical of fully driverless cars (Level 5) doesn’t necessarily equate to an unwillingness on the public’s part to take the next step (to Level 3 or 4). Gaining acceptance is a marathon, not a sprint.
Traveling into the future
I know I’ve just said, “let’s not look too far ahead” but this is worth a peek. Last week, Toronto-based Steve Lee of Aprilli Design Studio won the Radical Innovation Award 2018 with his “Transpitality” concept – integrating hospitality and transportation “into one form”. The Autonomous Travel Suite (ATS) is essentially a hotel room on wheels – and a pretty luxurious one at that (video here).
It’s nothing new; rather another vision built on the premise of getting away from the idea of conventional cars. The interesting thing for me is the use case. It’s designed with domestic travel in mind: that 6-10 hour drive for example. Often when faced with this prospect, especially in business, we opt for carbon-heavy air travel. Ten hours is just too long to sit in a car and be unproductive. But if it’s not a car you’re sitting in, and there is the opportunity for productivity as well as rest and relaxation, I for one could easily reconsider. In theory, this is an industry game-changer. In practice, it won’t be happening for a while.
So long, drive safely (until cars are driverless),