Let’s take a journey into a time when the vision has become reality – cars are driving autonomously on our roads. People trust the technology. What happened on the way there? This is the question asked by the beyond Initiative founded by Audi, bringing together experts from around the world. Their mission is to use artificial intelligence in ways that benefits society – such as autonomous driving and beyond.
Text: Lisa Marie Feldmann
Photos: Ulrike Myrzik
The time travel begins in the creative spaces of Audi’s Technical Development. Here, on this Thursday morning, above the roof of the Ingolstadt headquarters, philosophers are meeting with software engineers, and psychologists with lawyers. Between orange sofas and green stools, these advance thinkers are formulating hypotheses, clustering Post-its on a flipchart and processing questionnaires. In their minds eye, the world’s cars are already driving autonomously. And they are addressing the matter of the social framework necessary for this. “The central hurdle is social acceptance,” says Kate Darling, a robot ethicist from MIT Media Lab in Boston. “Because people will feel far more emotional about an accident caused by a machine rather than a person.”
Kate Darling is one of twelve international scientists participating in the workshop on the ethical, legal and social aspects of autonomous driving. They are in discussion with seven specialists from Audi, including software engineers, lawyers and policy experts. This meeting of experts is part of the beyond Initiative established by Audi to address the social impact of artificial intelligence (AI). The background to this is the conviction that AI can substantially improve our mobility, our world of work and our lives. To achieve this aim, science, commerce, politics and society have to work closely together. It is for this reason that, two years ago, Audi established this interdisciplinary network of international experts. One of the initiative’s focus topics is, of course, automated and autonomous driving.
The central hurdle is social acceptance.
Kate Darling, robot ethicist from MIT Media Lab
AI is a key technology for this. It helps the car to perceive and interpret its environment and to take decisions. “Automated and autonomous driving will deliver more comfort and convenience,” says Miklos Kiss, Audi’s Head of Advance Development Automated Driving and a member of the beyond network. “And it will make traffic safer.” Today, ninety percent of all accidents are caused by human error. AI in vehicles promises to reduce the number of road traffic accidents significantly.
Nevertheless, there are ethical concerns. The best-known example is a critical traffic situation where an accident is unavoidable – the so-called dilemma situation. In this example, an autonomous car is faced with three options – it either steers left and hits person A or it steers right and hits person B. If it drives straight ahead, it injures its own occupants. At the beyond workshop, the Audi experts discuss this problem with, among others, Iyad Rahwan from MIT Media Lab in Boston. He and two colleagues developed the online tool Moral Machine, which presents users with different variations of the dilemma situation. They then decide, by clicking the mouse, what the autonomous car should do. In real-life, a person would not have any time to make a considered decision. They would act instinctively. However, we expect machines to make the right decision.
Automated and autonomous driving will deliver more comfort and convenience.
Miklos Kiss, Head of Advance Development Automated Driving at Audi and a member of the beyond network
“It’s highly unlikely that the situation involving the decision between person A, person B and an individual’s own life would ever arise in exactly this format,” says Miklos Kiss from Audi. And obviously, automakers are doing everything to avoid such a situation. “All the same, the dilemma debate is good food for thought when it comes to thinking about other risk scenarios.” For instance, when it isn’t completely clear what will happen if a car steers in a particular direction. Is it ethically justifiable to opt for the unknown? Society is going to have to find answers to that.
In the legal context, public focus is primarily on who is responsible in the event of a possible accident. Right now, the law always applies to the individual. It seeks out, in effect, the person behind the machine. Within the context of increasingly intelligent machines, the long-term question, however, is whether perhaps there will eventually have to be a so-called “e-person”. This means that a thing will be attributed with a legal personality with independent liability. But it goes much farther than that – permission, data protection, road traffic safety, criminal law, constitutional law – autonomous driving touches a wide range of legal areas. “Social acceptance and legal frameworks; those are the two forces that – once they exist – will make driverless cars a real prospect, with all the associated benefits,” says Luciano Floridi, an information ethicist at Oxford University and participant in the beyond workshop.
Social acceptance and legal frameworks; those are the two forces that – once they exist – will make driverless cars a real prospect, with all the associated benefits.
Luciano Floridi, information ethicist at Oxford University and participant in the beyond workshop
Anticipating today the challenges of tomorrow – that’s the objective of the beyond Initiative. Besides autonomous driving, it also addresses another field of interest that will be changed considerably by artificial intelligence – our world of work. “We have to sensitize the public in this area,” says Martin Ford, author of bestseller Rise of the Robots. “They have to understand what’s happening here.” Because AI will have an impact on the entire value chain within companies and therefore on their employees.
At Audi, the transition is already evident. In Production, people are increasingly working hand-in-hand with robots. In the offices, intelligent algorithms are helping with the analysis of massive amounts of data. “Our aim is intelligent, networked production, in which people and machines work perfectly with one another,” says Tarek Mashhour, Head of Production Strategy at Audi. “For us, the human being is at the heart of the Smart Factory.” One example of the Smart Factory is so-called modular assembly. There will no longer be any production lines. Instead, bodyshells will be brought from workstation to workstation on self-driving transport systems. A self-learning algorithm ensures that all processes run efficiently and adapted to the employees.
Create acceptance. Build expertise. Initiate partnerships. Those are the objectives of the beyond Initiative. “It is very important to discuss these issues at an early stage and to try to solve some of the problems before they arise,” reckons Kate Darling from MIT Media Lab. The beyond Initiative is convinced that the change AI will make to our lives depends on all of us – and on how we use the potential of the new technologies.