Table of Content Executive Summary 3 Introduction 7 The ANE Hackathon 10 Operational Definitions 12 Pressing Issues for an Ethics of AI 17 Opportunities for addressing pressing issues 25 Recommendations and Guidelines 29 References 34 Further reading 37

Executive Summary The Global Risks Report 2017, published by the World Economic Forum, lists artificial intelligence (AI) and robotics as the third of twelve key emerging technologies. These emerging autonomous technologies are disrupting established business models, and changing society in ways that are not always easy to anticipate. The European Commission has stated that “the way we approach AI will define the world we live in”, and this is supported by the proliferation of discussion and concerns among policy makers, business interests and general public alike. While advances in machine intelligence provide unprecedented opportunities, large-scale data collection to support such opportunities represents significant causes for concern. To mitigate the risks, a regulatory framework is needed that includes ethical standards, normative expectations, assessment of responsibility, and accountability for actions. Questions such as who should take moral, ethical and legal responsibility for artificial intelligence technologies need to be prioritized on the political agenda. The Nordic countries are known for low levels of corruption, high levels of involvement of civil society in policy making and a commitment to ethical treatment of consumers and of the labor force. As such, the Nordic countries are positioned well to be frontrunners in setting the agenda for how to address the issues of ethics in AI development and implementation. Engineering plays an essential role in building, sustaining, and improving the quality of life for individuals in contemporary societies. In this way engineers are at the forefront of developing autonomous systems and adding machine intelligence to existing mechanisms and processes. The many standards and codes of conduct agree that one of the major responsibilities of engineers is to promote positive outcomes for society, and to limit harm. However, in a rapidly changing world, what comes to constitute a positive outcome, and what could potentially cause harm have become much more difficult to recognize. Current guidelines and standards often do not fully address the problems that engineers face and the responsibilities they must take on in working with AI. As the stakes rise so does the need for addressing the ethics of engineering in practice more directly. The Association of Nordic Engineers (ANE) membership is composed of national unions, all of which have their own versions of guidelines and codes of conduct for engineers to use in their work today. However, recent developments in AI and machine learning have outpaced most of the existing ethical guidelines and frameworks for best practices. In this time of global digital data economy and an increasingly fast pace of technological change, ANE sees the development of an overall position for all Nordic engineers on what constitutes ethical conduct with respect to AI as a crucial step forward. 3

    ...