Artificial intelligence (AI) is increasingly adopted in society, creating numerous opportunities but at the same time posing ethical challenges. Many of these are familiar, such as issues of fairness, responsibility, and privacy, but are presented in a new and challenging guise due to our limited ability to steer and predict the outputs of AI systems. This chapter first introduces these ethical challenges, stressing that overviews of values are a good starting point but often fail to suffice due to the context-sensitivity of ethical challenges. Second, this chapter discusses methods to tackle these challenges. Main ethical theories (such as virtue ethics, consequentialism, and deontology) are shown to provide a starting point, but often lack the details needed for actionable AI ethics. Instead, we argue that mid-level philosophical theories coupled to design-approaches such as “design for values”, together with interdisciplinary working methods, offer the best way forward. The chapter aims to show how these approaches can lead to an ethics of AI that is actionable and that can be proactively integrated in the design of AI systems.

All successful technologies change our lives. Up until the last fifteen years, cars had changed things more than computers had. Mainframe computers by then had changed administration and management, production in corporations, and scientific research, but they had a minimal effect on everyday life. It was really only with the advent of the World Wide Web and the incorporation of computer chips in many common appliances that the lives of most people were changed by computer technology. One of the most important features of information technology (IT) today is its ubiquity. This ubiquity is a result of what James Moor calls the logical malleability of computers. Computers can be programmed to do a large variety of different things; route information packets on the Internet, simulate hurricanes, make music, and instruct robots. They can be adapted to many different devices and put to many different uses. They allow us to work online, shop online, relax by playing computer games interactively with people from all over the world, get our news, study for our degrees, and find most of the information that we require.

The technology has not only changed what we do, but also how we do it. E-mail, chat rooms, blogs, and other forms of computer-mediated communication have altered how we communicate, and with whom we communicate and interact. It has changed how we relate to each other and how we experience our relations with others.

INTRODUCTION

Distributive justice in contemporary information societies concerns, among other issues, the distribution of information, information services, and information infrastructures. There is currently much enthusiasm about the potential of new information and communication technologies (ICTs), and the immense information resources contained on the World Wide Web, to improve the lives of those who have access to them. But at the same time, there is mounting concern over the uneven distribution of the new information wealth, both within nations and internationally. Some even fear that we may be heading for a regime of local and global ‘information-apartheid’ (Ratan 1995; Wresch 1996; Loader 1998; Schiller 1996), in which the ‘information-haves’ are segregated from the ‘information have-nots’.

A host of empirical research adds force to such concerns. According to a 2005 UNCTAD study (UNCTAD 2006) the digital divide in the first years of the twenty-first century between nations is still wide. A person in a high-income country is more than twenty-two times more likely to be an Internet user than one in a low-income nation. In a high-income country, Internet affordability relative to income is more than150 times better than in a low-income nation. Even in lower-middle-income countries, the cost of twenty hours of inferior Internet service is nearly one-third the average monthly income. In high-income countries the cost of Internet service is affordable for most households, although the report claims that, even within these countries, considerable digital divides continue to exist between urban and rural areas, genders, age groups, and racial groups.

Information technology allows us to generate, store, and process huge quantities of data. Search engines, satellites, sensor networks, scientists, security agencies, marketers, and database managers are processing terabytes of data per day. A good part of these data are about persons – about their characteristics, their thoughts, their movements, behaviour, communications, and preferences – or they can be used to produce such data. All countries and cultures in the present and past have constrained access to certain types of personal data in some way or the other (see Moore 1984). There are etiquettes, customs, artefacts, technologies, or laws, and combinations thereof, which prevent or proscribe against the use or dissemination of personal information. Walls, curtains, doors, veils, sealed envelopes, sunglasses, clothes, locked cabinets, privacy laws, secure databases, cryptographic keys, and passwords all serve the purpose of preventing individual persons to acquire and use information about other persons. The issues are often discussed in the context of a specific social sector or professional domain, such as health care, social or homeland security, search engines, marketing, or policing. More specifically the issues can be concerned with camera surveillance, the monitoring of Internet communications, the retention of Internet traffic data, the disclosure of passenger lists to security agencies, the availability of individual medical information in the public health system, the linking and matching of databases in social security to detect fraud, and sifting and trawling through financial databases in order to find suspect transactions.