More safety, less congestion: Artificial intelligence could revolutionize our mobility. But networking vehicles and infrastructure poses new challenges for IT security and data protection.
So far, Germany knows autonomous vehicles only as test objects. BMW, Audi, and Daimler are experimenting with autonomous driving on the digital test track on highway A9. In Hamburg, Volkswagen sent five highly automated electric Golf cars on a test drive downtown. BMW has been testing similar vehicles in Munich since 2017.
The small town of Monheim in North Rhine-Westphalia is now taking the plunge: This autumn, it is set to launch the first regular transit service with five autonomous electric buses. For now, they may not exceed 20 km/h (12mph) and each bus has a driver who can take control in an emergency.
The global race for autonomous mobility is fierce: In California, 62 automobile and technology companies from all over the world are testing autonomous driving. The Chinese government wants at least half of all new cars to be equipped with artificial intelligence (AI) by 2020.
Use of AI in traffic holds great promise: Proponents argue that it could make our cities safer and cleaner. 94 percent of all traffic accidents are due to human error. Autonomous vehicles, the argument goes, don’t drive under the influence, don’t get tired, and don’t get distracted when they communicate.
Digital traffic guidance systems can help avoid traffic jams and emissions. In Hangzhou, China, the technology group Alibaba has connected traffic lights, public transportation, and even police and fire departments in a “City Brain.” Between 2016 and 2018, Hangzhou dropped from 5th to 57th place in the ranking of China’s most congested cities.
Mobility experts also see great potential in private ridesharing concepts, which would reduce the number of cars on our roads. Uber, Lyft, and their Chinese counterpart Didi Chuxing are running tests with autonomous taxis in several American and Chinese cities.
Does all that sound too good to be true? That’s exactly what sceptics fear. Autonomous vehicles also make mistakes – just this spring, the autopilot in a Tesla caused a fatal accident in Florida. It is quite uncertain whether autonomous private vehicles would actually improve our CO2 balance. Drivers who can’t find a parking space could simply get out of the car and let the vehicle continue to cruise around - the result would be more traffic, not less. An expert study by the Berlin Social Science Center commissioned by the Green Party in the Hamburg City Council cautions that autonomous driving will only reduce emissions and increase safety if autonomous vehicle fleets are integrated in an intelligent traffic concept. The study concludes: “The opportunities of a digital traffic revolution cannot unfold without political control.”
IT security, consumer protection, and privacy
Another challenge is the IT security of networked vehicles. In today’s models, the infotainment system is often physically integrated with the control elements. In 2015, two security experts were able to access the diagnostic interface (CAN bus) of certain vehicle models via the internet. They remotely controlled brakes, acceleration, and other functions of a Jeep Cherokee.
Increasingly connected vehicles are also a problem from a consumer’s point of view. How long will manufacturers provide software updates for their vehicles before they discontinue security updates for older models?
One of the most complex issues is privacy. Autonomous cars are often referred to as “smartphones on wheels.” Connected driving involves the collection and processing of unprecedented amounts of personal and location data. Who owns this data? For which purposes can, should, or may it be stored and used?
Legislation on this issue is still in its infancy. Germany regulated automated driving in 2017 in its Road Traffic Act - but it has yet to lay down clear rules for manufacturers’ liability. The EU is debating an update to its C-ITS Directive (Directive on the framework for the deployment of Intelligent Transport Systems in the field of road transport and for interfaces with other modes of transport) to regulate the sharing of data between vehicles (Car2Car) and the vehicle’s communication with its environment (Car2X).
Trying to set global rules is even more complicated. In China, data from networked cities also feed into the “social credit system,” the government’s tool to perfect its control of society.
Western democracies must find ways to allow their citizens to retain control over their data. The German city of Duisburg wants to let Chinese IT group Huawei turn it into a “Smart City,” but it is facing uncomfortable questions about the project’s IT security.
Technical solutions include the anonymization of data or encryption of vehicle communications. The EU’s General Data Protection Regulation restricts data collection to such data that are actually necessary to provide a service (purpose limitation and data minimization). And these rules apply not only to Google and Facebook, but also to Daimler and BMW.