Ten years on from the blogpost that triggered the AV revolution
Today is exactly ten years since the Google self-driving car project first officially broke cover, in a blogpost authored by Sebastian Thrun and entitled ‘What we’re driving at’. Of course, automated vehicles (AVs) were nothing new at this time. Various concepts and trials, including the DARPA Grand challenges, had provided evidence that technologies could imitate human driving under certain controlled conditions and in particular environments. However, the confident swagger of this blogpost, the claim of 140,000 miles of automated driving on public roads and with Google’s intellectual, technological and financial horsepower under the bonnet, it seemed things were going to get very real, very quickly. I believe this was the trigger for the rush to develop automated vehicles that we have seen over the last ten years as start-ups, vehicle manufacturers, tier-one providers, technology companies and even countries scrambled to get ahead in the ‘race for autonomy’.
The post itself signposts many of the benefits of automation that many proponents of this technology quote today; in particular, the suggestions that automation technologies could reduce road fatalities by 50%, reduce car usage, cut energy consumption and increase productivity. Ten years later, they have not had an impact on any of those metrics.
Video from my trip to Google X in 2016
It is also five years since Google first completed a fully driverless ride on public roads with a single passenger, carrying the retired director of the Santa Clara Valley Blind Center, Steve Mahan - who had lost his sight years before. Closing his TED talk that same year, then head of the project, Chris Urmson, joked that they working so that AV technologies would mean his 11-and-a-half year-old son didn’t need to get this driver’s licence for which he would be eligible four and a half years later.
And, in a remarkable celestial alignment, yesterday, Waymo – formerly the Google self-driving car project – announced that their Waymo One service was being opened to the general public - almost ten years to the day after the original blogpost. Make no mistake, this is a landmark moment. Now we get to see if the promise of automated vehicles can withstand the challenges of public service. Making profit from moving people is not easy – and I’m sure Waymo is a long way from doing that – but we can now start to see if the promises about safety, efficiency, inclusivity and equity stand up to scrutiny. We’ll also get genuine feedback; whereas Waymo One riders to date were covered by NDAs, public riders will no doubt be keen to share their views. It’s an indication of Waymo’s confidence in the safety and performance of the system that they can make this step.
In considering these developments, I have three immediate reactions. Firstly, wow! The ability to make a passenger car drive on public roads in commercial service and without human supervision is an incredible achievement. Ok, the hype ramped up way too far and way too fast but now we’re opening a new chapter in the journey to automated transport. Following the terminology of the Gartner hype cycle, maybe we’ve turned a corner out of the trough of disillusionment and have started making the first tentative steps in the foothills up the slope of enlightenment. We can start to assess automated driving based on real world safety and productivity and not just promises of such.
Have AVs turned a corner on the hype cycle?
Secondly, a note of concern. With huge investment and their reputation at stake, I firmly believe that Waymo will have satisfied themselves that their systems are sufficiently safe to stand up to any scrutiny should there be an incident involving their vehicles. However, that is only a belief. It would be better if there was an agreed process by which those seeking operate such vehicles could demonstrate that they performed to an acceptable level in an independently verifiable process. This was the first recommendation in the Ethics of Connected and Automated Vehicles report that I helped to write as a member of an expert panel recruited by the European Commission and was recognised as a priority in the CAV standards roadmap that I co-authored for BSI as part of their CCAV-sponsored programme of work on CAVs.
And thirdly, I think about Chris Urmson’s son and the sense that, as soon as he reached the required age, he would need to get a licence to drive. So AV technology has not matured to the extent that he is able to be driven wherever he needs to go but, in the ten years since the original Google blogpost, perhaps AVs have had a more important impact. The potential radical transformation that AVs promised gave us the scope to think differently about how we want transportation to be achieved and ultimately, how we want to live our lives. This has only been further reinforced by the changes we’ve seen as a result of the COVID-19 pandemic as we see neighbourhoods (such as Brixton) seeking to retain the some of the changes that saw traffic reduce, air quality improve and walking and cycling feel more comfortable. With good planning, suitable infrastructure and efficient public transport, a driving licence shouldn’t be necessary in order to achieve economic, social, health or educational objectives. If AVs can contribute to achieving that in the next ten years, it will truly be a cause for celebration.