28 Feb 2017
Gillian Ku looks at the potential risks and challenges facing the negotiators
28 Nov 2016
Freek Vermeulen on self-driving cars, knock-on industry effects and what could happen to drivers’ skills
Driverless cars will be safer than cars driven by humans. At least, that is what some of the world’s leading experts are claiming. Elon Musk declared in October 2016 that critics of driverless cars are “killing people” as he announced Tesla’s intention to launch a fleet of its own. The CEO of Tesla and SpaceX pointed out that while fatal crashes such as the one involving a Tesla Model S in May 2016 make headlines, “1.2 million people die every year in manual crashes”.
Automakers and technology firms are responding to demands for safer cars by entering a race to put self-driving cars safely on the road. Ford, Tesla, Nissan, Google and more have announced their intentions to provide driverless cars for commercial use within five years.
“You can think of driverless cars as a substitute technology,” says Freek Vermeulen, London Business School Term Associate Professor of Strategy and Entrepreneurship. Such technology is a customer’s preference of one product over another due to its technological superiority. “We occasionally find a new technology that is an improvement in nearly every way and it almost entirely replaces the old technology.”
He draws a parallel with the gradual replacement of horses and carriages after the introduction of motor cars in 1893. “Just as people formerly thought of cars as horseless carts, driverless cars will become commonplace. Today, driving horse-drawn carriages still exists for hobbyists and I imagine driving cars could go down the same route.”
Dr Vermeulen explains that the journey to commercialisation involves legitimising the substitute technology. “It’s an unstoppable force. But de-legitimising events might delay the technology, as we've seen with Tesla and the autopilot function. Malfunctions are inevitable in the early testing phases of any product.”
The uptake of driverless cars is likely to follow an ‘S curve’, he says. On this curve, a technology’s performance evolves slowly at first but then has a breakthrough and performance improves rapidly.
“At first, car companies market their new products to specific segments. Take for example Tesla, which targets affluent people wanting a cool car. This is simply a strategy to gain legitimacy. After the car has become popular with that particular segment, Tesla can then launch affordable models for the mass market.”
The familiar car manufacturers of the last century – such as BMW and Volkswagen – will not necessarily be the dealers of the future. Competition is likely to extend to technology giants such as Google and Apple. Dr Vermeulen says, however, that the world often overestimates the speed of technological developments. “I remember that 10 years ago futurists said we would all be holidaying on the moon. It hasn’t happened yet.
“When dramatic industry changes take place there’s often a shift in power to where the most money is made. So even when people are no longer driving cars, there will still be elements of the car manufacturing process that will remain the same – the value will perhaps just shift to software.”
He takes the PC industry as an example. “Power used to sit with companies such as IBM but gradually value moved to operating systems like Windows. Microsoft, originally a supplier of IBM, steadily assumed power.”
So what are car manufacturers doing about it? “They’re aware of and responding to this risk in different ways,” says Dr Vermeulen. Renault-Nissan CEO Carlos Ghosn, for example, announced in October 2016 that his company had joined forces with Microsoft to deliver 10 self-driving cars by 2020 in an interview with TechCrunch. “Tech companies and car makers can join forces, each one doing its job,” he said. “I’m not trying to be a tech company, and the tech company is not trying to be a car maker.” He admitted that Microsoft provides valuable expertise in vehicle software.
Driverless cars have more access to data than human drivers, allowing them to make better-informed decisions. For instance, they are designed with sensors that offer a 360-degree view around the vehicle, enabling them to avoid hazards with much greater accuracy than people. But what will this high level of autonomy mean for future drivers?
Research into pilots’ skills in the automated cockpit reveals insight, says Dr Vermeulen. “As planes have become more automated over the last 10–20 years, pilots’ abilities have narrowed.” The ‘Retention of Manual Flying Skills in the Automated Cockpit’ study indicates that pilots who rely too much on automation can lose the critical thinking skills that enable them to adapt and deal with unexpected situations.
“This has two implications,” he says. “One, do pilots spot irregular readings or algorithm blips on instruments? Two, do they still have the skills to do something about it?” He adds that simulator-based training further amplifies such concerns. “People need to train in real life to deal with complex situations.
“We assume that drivers will be able to intervene when required and take over from the software in complex situations, but that’s going to be difficult when drivers are no longer paying attention. I doubt that operators of driverless cars will have enough experience of manual driving in the future – and that is going to be one of the big challenges.”