DAVID FURLONGER: Who’s really in control?

Many years ago, when I worked on a Scottish local newspaper, two of my colleagues got so drunk while reporting a far-flung village festival that after their car ended in a ditch on the way back to the office, police arrived to find them sitting dazedly next to the vehicle, arguing over who had been behind the wheel. Each was convinced that he had been the passenger and the other the driver. 

It’s a bit like being in a self-driving car. When you are at the wheel, who’s in charge, you or the vehicle? I’ve travelled in such vehicles several times, albeit under very controlled test circumstances. Even with no other vehicles in sight, it’s very hard to resist the temptation to grab the steering wheel. 

Maybe this sense of uncertainty is one of the reasons self-driving cars are taking so long to be accepted. Had original predictions been correct, there would be tens, perhaps hundreds, of thousands of them on the world’s roads by now, safely chauffeuring their occupants from A to Z and all points in between. 

Not in South Africa, of course. For self-driving cars to succeed, roads infrastructure must be well maintained and driver behaviour predictable. Between potholes, load-shedding, broken robots, drunken pedestrians, minibus taxis, lack of traffic policing and general driver anarchy, that’s a long way off. 

Self-driving vehicles are sometimes referred to as autonomous vehicles, though that’s a bit misleading. Most cars and trucks on the roads have a degree of autonomy. ABS braking was an early example of technology that can override the driver in an emergency. Today’s cars are full of safety features to keep  you a safe distance from the vehicle in front, avoid pedestrians and animals, stop you leaving your lane and even take over completely if you start to doze behind the wheel. 

The theory is that a “thinking” car is safer than an idiot human and will reduce accidents and road deaths. In many ways, this automotive technology was an early example of artificial intelligence (AI). 

The testing of self-driving (fully autonomous) vehicles is going on all around the world. Much of it is successful. However, as always with human nature, it takes a single negative headline to convince us that we’re all going to die in flames if we climb inside one. 

As international automotive consultancy S&P Global reports: ”One crash or fatality from one (autonomous) company can change perception of the entire industry.” 

Overcoming this negative image is proving harder than expected. Safety may be a prime motivator behind autonomous development but how does one measure it? Until autonomous vehicles match the billions of kilometres driven annually by humans, clear conclusions can’t be reached. “Simulation and extrapolation are indicative but not conclusive,” says S&P. 

Nor does it help that, despite using public roads to test their vehicles, autonomous developers are selective with the information they make public. “Companies are incentivised to share only the best information.” 

The fact that autonomous development is years behind schedule is not down only to perception. Companies were ridiculously optimistic about the timeframe needed to perfect self-driving. All kinds of unforeseen technological obstacles have got in the way. Put simply, companies underestimated the complexity involved. 

That’s why many hope driverless robotaxis will provide them with the market impetus and public confidence they crave. Mostly running on fixed routes, these vehicles offer the best initial hope for the manufacturing scale required to make them commercially viable. 

S&P predicts that China will eventually take the lead in the development and use of robotaxis, overtaking the US. That’s bad news for people already worried that China is primed for domination of the electric-vehicle market through its carefully planned control and exploitation of minerals used in batteries. 

Research shows that Chinese consumers are more enthusiastic about autonomous driving features and full self-driving than those in any other country, followed by people in India, Thailand, Brazil and Japan.  

Even now, though, the testing of robotaxis – some with humans on board in case of emergency – remains limited. In the US state of California, which prides itself on being at the forefront of everything technological, only 1,533 hit the road in 2022 – a 28% increase from the previous year. The sum total of travel by these vehicles was just over 8-million km. Brands at the forefront were Chevrolet, Chrysler, Jaguar, Toyota and Hyundai.  

China is less forthcoming about the scale of its testing. However, it’s worth noting that, according to S&P, the most popular US robotaxi models have 40 safety sensors on board and none has fewer than 21. In China, one model has 37 and the rest range from 15 to 20. 

How soon will robotaxis become commonplace? S&P states: “They have the potential to revolutionise mobility in the long term and displace some vehicle ownership. But they are still working through the early stages of commercialisation.” 

Another 10-12 years? That sounds about right, says the company.