In this article, I speculate that the nascent field of AI ethics might have a thing or two to learn from its older sibling, bioethics, and that field’s successes in establishing ethical reflection into the culture and practices of the medical profession.
In four important ways, the field of AI ethics today is not altogether different from the state medical ethics was in before it came into its full-fledged form in the 1970s. First, both fields had recently undergone exponential growth and advancement that had raised important new moral questions that society and the profession had not yet grappled with. Second, in response to these questions, both fields have drawn scholars from a wide array of traditional disciplines including theology, philosophy, sociology, medicine, and the law. Third, however, in their infancy, work in each field was often siloed into these disciplinary boundaries, limiting the import of the research. Finally, both medicine in the 1960s and the tech industry today seemed to lack the resources and culture to include ethical considerations as a regular practice of the industry. In the course of the next two decades the founding generation of bioethicists both convinced the medical industry that ethical reflection needed to be a central aspect of how medicine is done, and established the structures of a field of practical ethics.
A Very Brief History of Bioethics
Before medical ethics established an identity of its own, the work of bioethics was often being done in isolation among the far-flung corners of each parent discipline. Thus, it was likely that an article written for a theology or philosophy journal would be read by few medical professionals, while a case written up in a medical journal, unless it became a subject of popular media, never gained the attention (or the corresponding ethical rigor) of philosophical or theological ethicists. This began to change in 1969, when the Institute of Ethics, Society, and the Life Sciences (now known as the Hastings Center) became the first academic center in the world to be focused on the questions of Medical Ethics. The Hastings Center would soon be followed by the Kennedy Institute of Ethics at Georgetown University in 1971 and the two centers would, together, become the clearinghouses for a growing body of ethical reflections on the ever-increasing advancements in medicine in the latter half of the 20th Century.
The need for medical ethics was already well established by the time these twin medical ethics centers came into existence. The Doctors’ Trial at the Nuremberg Tribunal in 1947 had raised global consciousness of the need for moral limits on medical research and experimentation, but the conversation would not yet expand to the morality of medicine itself. In the decades that followed Nuremberg, however, James Watson and Francis Crick would discover DNA in 1953; Joseph Murray would conduct the first Renal Transplant in 1954 and Belding Scribner would invent Hemodialysis in 1960; the FDA would approve the first oral contraceptives in 1960; Christiaan Barnard would conduct the first successful heart transplant in 1967; and the U.S. Supreme Court would legalize abortion in 1973.
Each of these expansions of what medicine could do raised questions about what medicine should do. Responses first emerged from within the standards of medicine itself. For example, in 1968, the New England Journal of Medicine first defined an “irreversible coma” or what came to be known as “brain death,” thus redefining the moral point at which further medical treatment was expected. But the depth of the moral questions emerging in medicine called for the work to be undertaken by a field of professional ethicists. It was into this lacuna that the Hastings Center and the Kennedy Institute were established and the field of medical ethics was born.
The new field of medical ethics quickly set about trying to organize and define its work. “From the Hastings Center came a steady stream of carefully considered opinions, worked out by small groups of scholars from different disciplines called together to discuss specific issues in detail. From the Kennedy Institute came a four-volume Encyclopedia of Bioethics (1975) and the annual publication of the Bibliography of Bioethics, which, in volumes of steadily increasing size, registered the growing literature in the field.”
In the 1970s, certain key cases drew national attention, and crystalized the need for the field. In 1972, the Tuskegee Syphilis Trials were revealed, in which hundreds of poor, uneducated, black American men had been subject to a four-decade long study on the advancement of syphilis in their bodies, even when antibiotics had been well-established as a cure for syphilis since 1940. 1975 saw the case of Karen Ann Quinlen, whose parents elected to remove life support after she was determined to be in a “persistent vegetative state.” And in 1978, Baby Louise Joy Brown became the first artificially conceived baby brought to term. For each of these cases, the field of medical ethics was there to professionally examine the contours of each case, to make sense of the moral questions for the public at large, and to offer the various moral solutions that needed to be weighed.
Even as these high profile cases caught the national attention and the new field was there to offer its analysis, the field itself was continuing to deepen its local influence on medicine. Throughout the 1970s, local medical ethics review boards were being established in hospitals throughout the country to review and deal with challenging moral cases on the local level. At the same time, bioethics centers and programs were also being established at universities to both train the growing field of medical ethicists and to provide local scholarly outposts that could continue to produce new research as medicine itself continued to expand its capacities. By the end of the decade, the new field had not only been established but had created a whole network of research, documentation, and exchange of information that operated at both a scholarly and a clinical level. In one short decade, medical ethics had revolutionized how moral questions in medicine were handled.
The Lessons For AI Ethics
As noted above, the field of AI ethics today is in a similar nascent state as bioethics was in the late 1960s. Just as the medical advancements of the 1950s and 60s created new moral issues for medicine, in recent years AI’s exponential advancements in natural language processing, biometric identification, social media and advertising targeting, medical diagnosis, banking, autonomous transport, and industrial robotics (to name a few) have not only pushed the boundaries of what AI can do, but they have raised new important questions about what AI should do. Unfortunately, though, as with early medical ethics, where these questions are being asked is in siloed conversations among academics and intellectuals rather than the wide, public debates that lead to better industry practices that we need. In light of this, I cannot help but wonder if the nascent field of AI ethics might be able to learn from the history and example of its older sibling, bioethics.
Today we take for granted that ethical reflection is a part of how medicine is practiced. Doctors study the ethical as well as the medical aspects of new technologies and procedures. Medical decisions are subject to ethical review as a matter of course, and most hospitals maintain a staff of ethicists to provide patient consultations for difficult medical decisions. But when bioethics came to the fore in the 1970s, none of these practices existed. It is conceivable that the medical advances that began in the 1950s and 60s might have prompted a pushback on such ethical incursions into the medical industry in the name of “medical progress.” Indeed, this very dichotomy has kept some areas of the medical industry from embracing the field of bioethics as an interlocutor.
Even if not perfect, it is undeniable that bioethics hold more sway on medicine than most disciplines of practical ethics have on their respective industries. It would be worth studying how this came to be, and what AI Ethics might learn about how to make ethical reflection a greater part of the tech industry.
For example, does AI ethics need a watershed moment like the revelations of the Tuskegee Syphilis trials to catalyze a national conversation on ethics necessary for our latest digital age? The applications of AI today are not short on scandals. The Cambridge Analytica scandal had focused a spotlight on the data gathered by tech companies and the lack of protections on individuals. The 737-Max crashes coupled with Tesla autopilot failures have raised serious questions about the limits of machine autonomy. Reports of biased training data in algorithms have called into question the use of AI in criminal sentencing, hiring processes, loan applications, and so forth.
Yet none of these has prompted the kind of Presidential Commission on Individual Rights in an AI world that would parallel the National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research. Why is this? Were our social norms and political system just very different in the 1960s, or is there more to it? In our attempts to make AI ethics relevant, are there lessons to be learned from how bioethics made itself relevant? Certainly, we should not invite a tragedy on the order of the Tuskegee Syphillis experiments that prompted the National Commission, but are we reacting to existing scandals in unproductive ways?
Furthermore, should AI ethics look to the role that the Hastings Center and Kennedy Institute played in the development of bioethics? Unlike 1960s bioethics, today’s field of tech ethics is not short on academic centers and more seem to be formed every year. But among this sea of tech ethicists and tech ethics centers, none seem to function as a central clearinghouse and organizing place for discussions within the field, as the Hastings Center and Kennedy Institute have for the field of bioethics since its infancy. Would industry pay more attention if the field was more organized and centralized?
Finally, and importantly for AI & Faith, is there an important role for theology to play in the future of AI Ethics? The bioethics revolution in the 1960s was deeply influenced by theologically trained ethicists such as Paul Ramsey, Al Jonsen, Karen Lebacqz, Robert Veatch, James Childress, Allen Verhey, and James Gustafson (to name but a few). Early bioethics seemed to understand that, while universal agreement on the moral standards of the industry was probably impossible, much could be gained by gathering an overlapping consensus around salient points and principles, and that such a consensus would only be valid if it included religious perspectives. Outside of certain circles, this kind of thinking simply is not found in AI Ethics, and in fact, the industry can be hostile to any theological perspective. So for those of us who see value in understanding AI through the lens of faith, is there even more we can learn from the history of bioethics?
In conclusion, my goal here in this piece was merely speculative and I am aware that I am leaving the reader with more questions than answers. More research needs to be done. Nevertheless, we humans often need to be reminded that we do not always have to reinvent the wheel. Thus, what could be gained if we in AI Ethics looked to our sibling field that has made more inroads into both public consciousness and industry practice than any other field of ethics before or since?