An Ethics Framework for Research into Heterogeneous Systems

Abstract

Heterogeneous systems often found in the Internet of Things (IoT) have a wide range of challenges in ethics and law. Any device with an IP address can potentially collect, process, store and share data and make automated decisions in unpredictable ways. When conducting research and development in IoT, it is necessary to have a comprehensive socio-technical understanding of decision-making and datahandling, as well as procedures in place to pre-empt and address unforeseen consequences. In this paper we propose a comprehensive conceptual-modelling approach to help researchers systematically identify, consider and respond to challenges in ethics and law when conducting research and development of heterogeneous systems. Our framework is a six-layered model that addresses these concerns with regards to proximity to the data and actions in question. Using our framework, researchers should be able to deliver use-case scenarios that should be peer-reviewed by a large number of experts in dissimilar domains with the aim of identifying issues to why the research and development proposed is not done responsibly, so researchers can address these concerns. Finally, we explore a IoT use-case scenario, and we propose future directions for this work.

Privacy-certification standards for extended-reality devices and services

“…Happa et al [6] make the assumption it is impossible to identify all ethical and legal issues that can emerge from IoT research, and propose that through aggressive peer-review of conceivable data processing and decision-making in novel technologies, it is possible to tackle ethical and legislative concerns proactively and reactively in more well-informed ways. Such a methodology could be applied within an XR privacy context as well.…”

Achieving Ethical Algorithmic Behaviour in the Internet of Things: A Review

“…A comprehensive framework to help researchers identify and take into account challenges in ethics and law when researching IoT technologies, or more generally, heterogeneous systems, is given in [103]. Review of research projects by an ethics review board, consideration of national/regulatory guidelines and regulatory frameworks, and wider community engagement are among the suggested actions.…”

Achieving Ethical Algorithmic Behaviour in the Internet-of-Things: a Review

“…As reported in the Technology Review on bias in natural language processing systems, 40 due to the use of machine learning to learn how to recognise speech, there are issues with minority population groups due to lack of training examples for the machine learning algorithms: "If there aren't enough examples of a particular accent or vernacular, then these systems may simply fail to understand you." The original intention and motive of developers could be considered when judging algorithmic bias, and care is needed to determine if and when bias does arise, even if not originally intended, especially with machine learning on data.…”

“…Other standards and guidelines for IoT data privacy and device security are also being proposed and developed. 68 A comprehensive framework to help researchers identify and take into account challenges in ethics and law when researching IoT technologies, or more generally, heterogeneous systems, is given in [40]. Review of research projects by an ethics review board, consideration of national/regulatory guidelines and regulatory frameworks, and wider community engagement are among the suggested actions.…”