Author: Aimee van Wynsberghe

In 2018 conversations about AI are filled with concern for jobs, lack of trust in companies using AI on their platforms, anxiety about the lack of privacy and control over the collection and use of personal data. As dark as this picture is there is hope that some governments and companies want to do more. Today, ethical issues related to AI play a strong role in the discussion: governments, NGOs, companies, and academics are all aware of and pushing for consideration of ethics as a central part of AI research and development.

Although the current focus on ethics and societal risks coming from AI is promising there is still some confusion about what ethics is and/or what value it might add. To date, ethics has a bad reputation that stems from the fact that ethical concerns are noticed after a technology has been made. Ethics, from this vantage point, is characterized as the finger wagging parent telling off a child for being irresponsible or not considering the possible repercussions of his/her actions. Ethical reflection, done in this way, is seen as a hindrance, stifling innovation, and failing to see the good in a technology. And of course one would have a negative image of ethics when this is all the ethicist is able to do – critique a technology after it’s been made. But the ethics of 2018 and beyond needs to be much more hands-on and engaged earlier on in the R & D process. Consideration for ethical issues must be made a priority from the beginning of the design process so that ethics isn’t an after-thought but is instead a part of the life cycle of product development.

What is Ethics?

A full account of ethics is not possible in this short piece; however, it is important to note that ethics is an ongoing activity – a process. It is not something that can be entirely captured in a code of conduct or general principles as it is dynamic and not static. I find that the most useful conception of ethics can be found in the ancient philosophers who wrote that ethics is the study of the good life – both what the good life is and how to achieve it. It concerns both the individual level of understanding what the good life is for yourself as well as fitting that understanding into the good of society. While those may seem at odds, one cannot forget that they often depend upon a well-ordered society to achieve their vision of the good life. Ethics looks at the values that are important in our lives and are constitutive of the good life: privacy, safety, security, health, wellbeing, dignity, freedom of movement and so on. Identifying these elements that lead to the good life and then thinking about the kinds of rules and practices we need in place to achieve them.

And where does technology fit into this discussion of the good life? There are several views on the relationship between ethics and tech. Some believe that technology is neutral and it is only how humans use the technology that determines if it is good or bad. And there are others who claim that technology is value-laden: it is the result of a process in which the values, assumptions, and biases of a designer or a design team find a place in the resulting technology. This does not imply that designers are malicious in their inclusion of values into a product; rather, this is more often than not an act of omission whereby designers are never questioned to make their values or choices explicit.

Beyond this there are also different ways of conceptualizing the relationship between technology and the good life. Some technologies might help us achieve components of the good life while other technologies might change the components of the good life. For the former, take close relationships with family and loved ones. This is a long standing component of the good life and yet in the globalized world of today many people live at a distance from their loved ones. Technologies like Skype, Google Hangouts, or WhatsApp help people to stay in contact with those at a distance. In this example, the technology hasn’t changed the requirement of family relationships but has helped many to achieve it. But to take another example, consider the pervasive use of smartphones, for many of us the portable computer that puts maps, restaurant searches, or online shopping at our fingertips has recently become a part of many people’s conception of the good life. The constant connectedness ready at hand has, for many, become a component of the good life.

With this in mind the question to consider is how will AI allow us to achieve various components of the good life? Of equal importance is the question, will AI create new components of the good life and if so which ones? It is these questions that must guide our future with AI; we must not wait for the pervasive use of AI to decide whether it has been for good or bad.

The Solution: Ethicist as Designer[1]

As an ethicist who has worked closely with people making technologies I can’t stress enough the importance of the role of ethics advisor or as I have referred to it elsewhere, the ethicist as designer. The ethicist as designer is a part of the design team. He/she is not the same as an ethics committee – you don’t submit an application for research to be approved or disapproved. The ethics advisor is someone whose job is to join the design team to educate and raises awareness about possible ethical concerns in tandem with the R & D of AI. More importantly the ethicist as designer is meant to stimulate the moral imagination of those making the technology: how could it be used for bad in the future, is there something we could do to prevent it from being used in a malicious way?

Some people have argued that this kind of reflection can be done by the engineer or designer themselves. I disagree. It is unfair and unsustainable to place that kind of responsibility on the engineer. An ethicist and an engineer undergo drastically different training, they are experts of a different language, they are a part of different research communities and keep up to date with vastly different literature and trends. It is unfair to expect the engineer/designer to be up to date with the current state of the ethics of technology whilst remaining experts in their own field. The same goes for the ethicist. We cannot expect the trained ethicist to learn to code and become a programmer. While she may be capable of learning some basic skills it would be unfair to require that he/she engage to be an expert in a new domain. Instead, the idea is for the ethicist’s expertise to be integrated into the design team within academia or industry in the same way that there is a legal expert, a data specialist, a graphic designer, etc. on many teams. If Universities and companies are going to commit to the inclusion of ethics within the design space it must be more than merely having a code of conduct or ethical principles to guide the institution; rather, it must be done in a way that accounts for the explosion of thought and nuance that is the ethics of AI.

Examples of the current push for ethics are the Partnership on AI, the Google Principles on AI, the Asilomar Principles, the DeepMind principles to name a few. But many of these initiatives will fall short because they don’t quite grasp the main tenants of ethics. First, ethical reflection must happen early on in the design process and second ethics must be understood as an ongoing process, as a continuous activity that becomes more robust, more nuanced alongside the R & D of AI.

Instead we need companies who commit to the development of ethical products from the earliest stages of development like the FairPhone company making the world’s first ethical and sustainable smartphone. Or initiatives like FairTrade and Tony’s chocolate who work to instil ethical values throughout the entire chain of development. These companies didn’t see ethics as stifling their innovation process; instead, ethics was the source of their innovation. These companies used the values of sustainability, justice, and fairness as inspiration for their development rather than as constraints for their business. We need to do the same for AI today – we need companies to stand up to the plate to ethically create AI driven products and to engage in responsible AI R & D. And the way to get there is to add the crucial role of the ethicist to the design team.

[1] For more on this see van Wynsberghe, A. & Robbins, S. Sci Eng Ethics (2014) 20: 947. https://doi.org/10.1007/s11948-013-9498-4