Interview
'Society too complex for EU's self-driving cars dream'
By Peter Teffer
A veteran tech expert who advises the EU commission has warned that its dream of self-driving cars will almost certainly never happen - despite the EU's apparent embrace of the concept.
Marleen Stikker is the founder of the Dutch research organisation Waag, and one of the Netherlands' pioneers in creating a digital society online in the early 1990s.
Join EUobserver today
Get the EU news that really matters
Instant access to all articles — and 20 years of archives. 14-day free trial.
Choose your plan
... or subscribe as a group
Already a member?
A conversation with Stikker about technology quickly turns into one about language itself and the "frames" and "narratives" that are used.
Take for example 'autonomous vehicle', or any 'autonomous' system.
According to Stikker there is nothing 'autonomous' about those machines or algorithms, owned by private corporations.
"There is not a single autonomous system in the world. Nothing is autonomous. That is a totally fictitious idea," she told EUobserver in an interview.
"When the word 'autonomous' is used, that disguises that companies own the software," said Stikker.
The immediate future of technology is often framed under the rhetoric of 'Will robots take over your job?' Stikker has a different take on this.
"No, a company takes over your job, through the robot. The companies pay and programme the robot," she said.
Stikker said the argument that 'technology is neutral' does not hold the same way it does for, for example, a knife.
"You can do a lot of things with a knife, so it depends on the intention of the user whether they will cut bread or stab someone," she noted.
"But you cannot say that about software. Software carries out instructions, and those instructions have been preprogrammed," she added.
Single market dominates
Stikker has sat on several EU expert groups giving advice to the European Commission. Most recently, she has been appointed to an expert group on innovating cities.
A recurring effort is to try to influence the narrative, to make sure the debate is not only about companies.
"The dominating narrative in the EU is that of the single market," said Stikker.
She tries to supplement that by arguing that Europe is not only a 'market', but also a 'commons'.
In an email exchange that took place after the interview, Stikker also emphasised that she was not against all companies.
"I differentiate between companies with an 'extractive' business model, which extract value without democratic supervision, and companies that are 'generative', retaining value, reciprocity and based on public values," she wrote.
What helps is that the European commission itself, and more broadly the EU, is not a uniform being.
"There are also people within the commission that want a different approach," she said in the interview.
Lobbying
But it is difficult, she admits.
"The lobbying power is obviously on the side of industry. There is no way that in the short term we will be able to counter that with the same amount of lobbying power," said Stikker.
This became apparent when EUobserver looked at how the commission developed its position on self-driving cars.
While privately, commission civil servants would say that fully autonomous vehicles driving in European cities are still a long while away, the EU's official publications say otherwise - not if, but when.
Stikker disagrees that it is inevitable that self-driving cars will flood the streets of Europe.
She dismissed the claim that autonomous cars will make fewer accidents than cars with human drivers.
"It is an assumption that if all cars are autonomous, we will have fewer accidents," she said.
"That is not a true statement, it is an assumption that cannot yet be falsified, because we do not have enough autonomous systems to falsify it," Stikker said, referring to Karl Popper's scientific concept of falsifiability.
"We have no idea yet. No one can prove it - at best [you can show] simulations in which you have all the actors under control."
Autistic cars
This is one of Stikker's key reasons why society in the end will not choose to adopt fully self-driving cars - in order for them to work, the entire public space, including pedestrians, dogs, lampposts - would have to adapt so that the machine can 'see' them.
Between drivers and pedestrians, there can be a dialogue. But that social dialogue is not possible in the same way with machines, explained Stikker, adding that autonomous systems by nature are "autistic" and "narcissistic".
"We would have to reduce ourselves to something the machine can understand," she said.
Another argument against Europe embracing fully autonomous vehicles, is that the transition period is near-impossible to implement.
"Even if [the fully autonomous car] is [technically] possible, you would still have a transition period of 20 to 25 years during which you would have a mix of systems, where in some parts of the city there are smart sensors, but in others there aren't," she noted.
From drivers to operators
And then there is the challenge of the 'semi-autonomous' car, that is, one in which humans still have a role.
"When you are driving, you are in a flow," she said.
But it is much more difficult for the human brain to suddenly take the wheel if you were not paying attention.
"Pilots and train drivers need long training to be able to act at the right moment," she said.
A good driver does not necessarily make a good operator, said Stikker.
"Not everyone can do that," she noted.
A switch to fully autonomous cars could maybe be foreseen in a totalitarian state, but not in a democracy.
"The complexity is too big for that. Fortunately. The world is not a computer model," said Stikker.
Site Section
Related stories
- How the EU commission got tunnel vision on self-driving cars
- Europe debates AI - but AI is already here
- EU in race to set global Artificial Intelligence ethics standards
- From Malta to Poland: each EU state to have AI strategy
- Wifi or 5G to connect EU cars? MEPs weigh in
- EU official: 'weak politicians' appease taxi drivers