So what happens when a dream turns into, well, a pipedream? I spent today to find out by attending a conference on autonomous vehicles organised by Zenzic, the government company which is supporting the driverless initiative with a cool £250m or so of taxpayers money.
Remember the dream. After the first edition of my short book Driverless cars: on a road to nowhere was published in 2018, I spoke at several conferences around the country on the subject. Invariably, on the platform, there was a proselytizer, a true believer, who presented a vision of the future which was usually just a few years away ‘which would change all our lives’ as the norm would become a transport system dominated by driverless electric vehicles that would be shared use. Suburban roads would be grassed over as no one would need to park on them any longer, motorways would become less congested as these clever cars would driver closer together and people on their way to work would be able to sit in the back seat while being chauffeur-driven there by a robot and the vehicle would then be despatched to take the kids to school.
Although initially I was impressed when I first read about driverless cars in the Evening Standard in 2016, as soon as I started examining the concept, I realised the whole thing was a con. None of the assumptions about how this world would come about were founded in reality. The technology was far more difficult to develop than had been envisaged, no one wanted to share their cars and dispense with their own, and none of the promises, such as greater safety and reduced congestion rang true.
At the time I was a bit of a lone voice but gradually as time wore on, the evangelists’ case began to fall apart. Big players like Uber and Apple pulled out despite spending billions of dollars, others like Argo AI admitted it was all too difficult and dissolved themselves, and the major pioneers such as Waymo’s Chris Urmison and the ubiquitous Andrew Lewandowski began to make clear that their initial predictions had been far too optimistic and that this brave new world was, frankly, not going to happen.
So at the conference, I wanted to test the waters and discover what the mood was now. I even asked a question to what happened to all those brave predictions, and it was instructive that even Professor Nick Reed, who has worked on various Autonomous Vehicle projects for many years admitted that it was not going to happen ‘in any of our lifetimes’ – and there were some young people there! In fact, the only real enthusiasm for the whole concept came from the speaker from the Business Department and even his slides betrayed the fact that the promise made by Boris Johnson’s administration that there would be driverless cars in 2025 on British roads was not going to happen.
Instead, we got a litany of obstacles. Peter Davies of Thales stressed the risk of cyber terrorists and the lack of understanding of the potential disastrous implications of a concerted attack which ‘could come from ‘anywhere in the world’. He also focussed on data which is a complex and difficult issue. Unless companies are prepared to share data about their technology and its performance, it will be very difficult to assess their safety records. Moreover, accident investigators need to find out the circumstances of accidents and this cannot happen unless companies are prepared to be transparent. Yet it is already clear that Tesla has no intention of allowing third parties to obtain relevant information about its systems.
Jessica Uguccioni from the Law Commission, despite being personally optimistic about the future of the technology, highlighted several difficult issues notably the question of when and if AVs should be able to break rules. She gave two instances: while driving or parking on the pavement is generally illegal, there might be times when an AV would have to take that option rather than, say, hitting a pedestrian or even a cat (she owns two cats!). Secondly, she cited what I have called the Holborn problem of pedestrians crowded out from busy pavements and wandering into the street. The idea that cars should be able to nudge pedestrians out of their way is unlikely to be socially acceptable but the alternative is potentially gridlock.
The worst and most disappointing speaker was the guy from Waymo, Trent Victor, its head of safety research and best practices. Not only was he a poor speaker, hesitant and inarticulate, but the examples he used to demonstrate the development of the technology were completely unconvincing. He showed a car driving through a busy San Francisco street , but only for a few hundred yards with no difficult manoeuvres and his example of how AVs would avoid an accident was even worse. He cited a case where a driver going 107mph through a red light had killed another motorist, and said that it the speeding car had been an AV, it would not have gone that fast or through the light. Sure, but that was hardly the point. What if the AV had been the other car – would it have avoided the accident – answer came there none as the dodged the question panel.
Apart from these specifics, what was most interesting was the lack of any belief in the model that only recently had been seen as the inevitable outcome of the development of the technology. Quite simply none of the speakers or the people I spoke to during lunch believes in it any more. Yes, there may be uses such as autonomous tractors and even possibly trucks on US Interstates, but that is a poor return for a technology which has cost tens, if not hundreds, of billions of dollars. The question overhanging the conference, and indeed the whole industry that has emerged, is ‘what is it all for?’ One of the people I met at lunch had a marketing background and she asked the right question: ‘Have they asked if anyone wants any of this?’ The answer is definitely no.