Privacy Protection, Ethics, Robustness and Regulatory Issues in Autonomous Systems

One of the most important challenges of the present decade in Autonomous Systems (AS) and CPS is the accommodation of ethics, security and privacy issues related to embedded intelligence. Drones and Autonomous Vehicles (AV) are multipurpose AS with civilian, police and military applications, thus their prototype design includes components that may be built for dual use purposes. Second, AS suffer from different cyber-attack types, and some degree of cybersecurity is required. Moreover, as ASs misuse can be accidental or deliberate, it may lead to safety risks, to security risks of both physical and virtual assets, and potential infringements of privacy. Unfortunately, there is no specific legislation that prescribes the protective measures to misuse avoidance and vulnerability exploitation of ASs. Thankfully, there are some technical measures that should be considered in the design stage, mitigating some of these risks. On the legal perspective, privacy laws have been examined to govern ASs usage. However, these regulations still do not govern issues related to what kind of data can be collected by an AS and what ASs owners can deal with these data. As drones and AVs collect footage data, raise privacy and security concerns, related to flying boundaries, data collected in public and private spaces, stored and disseminated data. An increasing number of studies tackle on privacy and security concerns, on effective use of geofences, designated spaces, as to reinforce privacy for users and security for ASs. Data management and protection of AVs collected data is still at a “nascent stage”, as there are still some unanswered issues: e.g., the distinguish between personal and non-personal data; capability of “re-identification” etc. This lecture overviews all these aspects and prescribes some technical solutions towards risk mitigation.