“Understanding vehicular automation by connecting technological and human challenges”

Droit et robots : Droit science-fictionnel et fictions du droit"  (Law and robots: Science fiction law and legal fictions) is an interdisciplinary work published by Presses Universitaires de Valenciennes, featuring a chapter by three researchers from Université Gustave Eiffel. This book is aimed at a broad audience and calls for greater awareness of the challenges connected to our world's digital transformation. Read on for more details from Mariana Netto, automation engineer at PICS-L and lead author of the chapter focusing on the robotisation and growing automation of vehicles.

What are the implications of vehicular automation for users’ driving? What influences will it have on society? Should we accept these systems, given the current research into liability in the event of accidents? What role(s) does the variable of time have, given that it has multiple dimensions (experience time, driving time, accident time, etc.)? These are just a few of the questions raised by Mariana Netto, Jean-Marie Burkhardt and Dominique Gruyer, researchers from Université Gustave Eiffel, and Andrea Martinesco, legal PhD student at Université de Versailles Saint-Quentin-en-Yvelines, in their chapter of over 30 pages in the book “Droit et robots : Droit science-fictionnel et fictions du droit”. “The chapter opens by diving into the fictional universe of automated technology in transport and drawing connections between various disciplines. It tackles the challenges of vehicular automation, specifically through the lens of varying levels of automation,1 ranging from driver assistance systems like the anti-lock braking system (ABS) to conditional and full automation, explains Mariana Netto, automation engineer at PICS-L.2It all depends on the level of danger in a given situation that the system is designed to act in, which translates as time remaining before an accident occurs. This will determine the system's choice of level of automation, and how much it intrudes to help the driver in the situation. Some systems, such as ABS, were made mandatory well before international treaties were amended to recognise new driver assistance systems arriving on the market, and that is because ABS operates with a very short timescale and acts as a complementary measure that does not interfere with the driver's actions. On the other hand, in automatic cars, there are more challenges in designing these systems, as models are highly non-linear. ”

Machines shouldn't change our values.

In their text, titled “Rising levels of vehicular automation, from manual driving to fully automated: interdisciplinary perspectives combining technological sciences, cognitive sciences, human factors and law”, the authors discuss challenges linked to drivers’ ability to “retake control of the vehicle, in the event of situations that automated systems cannot handle”. They also detail the technological and human challenges of designing and developing driver assistance systems, as well as the ethical questions linked to fully automated vehicles. “How can we make good use of technology's many advantages, while imposing the necessary limits? "How can we ensure that technology is designed responsibly?As someone who designs technologies, you have to ask yourself these questions,” considers Mariana Netto. “For example, I don't think that a vehicle should have to choose who to save or not - machines should not change our values. On the other hand, our role is to ensure that technologies are safe and respect humans, by reflecting well before actually designing them.

The book “Droit et robots : Droit science-fictionnel et fictions du droit” is edited by three researchers3 as part of the Robotic Laws and Human Avatars IdEX project. “The aim of this work is to raise awareness, especially among decision makers, of our world's digital transformation and the legal, ethical and social consequences of the arrival of robots, AI and augmented humans. It has also been created as a reference text for students.” The three researchers brought together other researchers and university professors in the fields of automation, literature, psychology and law for a day-long conference before contributing to the book. “It was an incredible, inspiring, interdisciplinary experience,” confirms Mariana Netto. “Gaining insight into other disciplines made me look at the technology I study with fresh eyes. Now, I aim to develop algorithms while questioning myself and anticipating potential impacts on the environment in which the automation system will evolve. This ethical, multidisciplinary approach to science is growing all around the world. And our university is part of this movement.

1 Society of Automotive Engineers (SAE) classification: level 0: the vehicle is not automated; 1: automated driver assistance functions (ABS, etc.); 2: partial automation of central functions (lane keeping assistance, etc.); 3: conditional driving automation; 4: essentially self-driving vehicle; 5: full automation in all circumstances.

2 Laboratory of Perceptions, Interactions, Behaviours and Simulations of Road Users, created by the merging of LIVIC and LEPSIS on 1 January 2020.

3 Frédérique Berrod, Philippe Clermont, Damien Trentesaux