Ethical Behaviour Aspects of Autonomous Intelligent Cyber-Physical Systems

Engineering ethical behaviors in autonomous industrial cyber-physical human systems

“…Several efforts focus on the applicability and engineering of ethics in autonomous systems (Arkin et al 2012;Westerlund 2020;Vanderelst and Winfield 2018). Special concern is placed upon I-CPHS since these should not be only safe to use, but also be trusted and ethical (Winfield et al 2014;Trentesaux and Karnouskos 2019). Overall, engineering ethics in autonomous systems is considered a pertinent challenge, for which, however, there are still several issues that are insufficiently addressed (Winfield et al 2019), and that although we are at the beginning of a mass introduction of autonomous systems in society, e.g., with self-driving cars (Karnouskos 2020).…”

“…More, sometimes, designers adopting a deontological approach do not realize that the rule they designed would let in fact remaining degrees of freedom for complementary decisions yet to be taken. For example, a deontological rule in train transportation "emergency stop in case of fire" contains hidden complementary decisions to be taken by the conductor: it does not mean "stop right now", which lets room to decide the exact moment to stop, especially if the train is on a high bridge or in a tunnel (which should be avoided as the evacuation of passengers would be risky) (Trentesaux and Karnouskos 2019). This can be extrapolated for the future autonomous train, with no more conductor.…”

“…A certain level of reluctance by researchers in engineering on this topic exists (Trentesaux and Karnouskos 2019), and, therefore, it is important to raise their awareness in the context of Industry 4.0 and I-CPHS. This need for awareness can be illustrated through the following example.…”

“…In that sense, safety can be seen as an additional dimension that complements machine ethics (Winfield et al 2014): machine ethics complements the concept of safety beyond its core scope, towards developing and integrating other dimensions, such as trustworthiness, data protection, and privacy, altruism, politeness, accountability, etc. (Trentesaux and Rault 2017b). For instance, the examples provided in Table 1 typically concern the dimension of altruism.…”

“…As such, societal acceptance of autonomous artificial systems (robots, selfdriving cars, etc.) is expected to remain a delicate issue, and so is also the proper engineering of ethics in such systems to lead to their societal acceptance (Trentesaux and Karnouskos 2019).…”