Bioethics Forum Essay
What If AI Could Be a Bioethics Scholar?
Twenty years ago, I wrote the first essay published in Hastings Bioethics Forum. At that time, online publications were new and many people were suspicious of them. My essay reflected on the reliability of the internet for scholars: Would the links to their publications remain?
“The other day I discovered that an article that I had read on the Internet no longer exists,” I wrote – not an uncommon experience back in 2006. “When I typed its URL the web page that opened announced simply, ‘File Not Found,’ in an intimidating large font.”
Recently, in recognition of the Forum’s 20th anniversary, I reflected on my reflections in that essay. Now that most publications are online, there is something quaint about being concerned about losing files. We seem to be in an age of too many files, thanks, in large measure, to AI churning out content. Amazon has limited the number of books that an author can publish on its site to three a day in response to a flood of AI-generated books.
The big question that is preoccupying me is about the implications of AI on scholars. Recognizing that researchers are using AI tools to prepare manuscripts for publication, academic publishers have developed guidelines, including the requirement of one publisher that authors may only use AI technology as a companion to their writing process, not a replacement. Such guidelines emphasize the importance of keeping humans in the loop. The editor of a neuroscience journal recently resigned in protest over the journal’s misuse of an AI system for finding reviewers that overrode his selection.
But what if AI large language models can one day create scholarship independently? How would we humans know?
I would like to propose a kind of Turing test for bioethics scholarship. To pass, AI would have to demonstrate that it could create new knowledge. AI tools can already produce papers that could be given a high grade in a college class, but the level of work at that point in one’s education tends to involve synthesis, not innovation. PhDs are expected to not simply produce work based on secondary sources but instead to produce new ideas.
There are two parts to my test that AI would have to pass to demonstrate that it created new knowledge and made a genuine contribution to the field of bioethics.
First, can it surprise us with what it creates? Consider, for example, the ways in which AI composes music. While the music may seem new and equivalent in listening experience to music created by humans, the compositions rely on prior musical creations in a particular genre. In the history of music, however, there are moments when artists broke significantly from established patterns. Think of John Cage and Ornette Coleman.
Scholarship also relies on novelty and innovation. I doubt that AI is capable of the innovative thinking needed to have, for example, questioned classifying deafness as a disability, used narrative theory in ethics analysis, or revived casuistry.
Second, can AI identify something in need of correction? This is an issue of central importance to bioethics. William May noted that ethics necessitates a “corrective vision,” that is, the ability not merely to see the world as it is but also the desire to imagine how it should be. In my Turing test for bioethics scholarship, a central focal point is whether an AI system can view the world as needing revision.
I know that I can give an AI tool a moral problem and it can resolve it using established methods in the field. As soon as an AI program became available to me, I gave it a clinical ethics case to analyze. The case involved a patient refusing a medical intervention that had a low risk and high benefit. Within seconds, the AI explained which bioethics principles were in conflict and the strengths and weaknesses of selecting one principle over the other. The program recommended that, if the patient has decision-making ability, the decision to refuse treatment should be respected.
The ability to find solutions, nevertheless, is often secondary to the task of seeing the world as needing correction: the ought in the is. If AI had been asked to provide ethics oversight to the Tuskegee syphilis study, would it have questioned the ethics of the project? Not without corrective vision.
Others may wish to add conditions for a bioethics Turing test, but I think that the notions of novelty and corrective vision are two essential components that such a system would have to demonstrate.
If AI can eventually produce genuine scholarship, what relationship will humans have with this work? Will AI come to consider us not only unnecessary to furthering ethical analysis but too slow and dull-witted – the intellectual equivalents of the lazy, helpless humans in the movie WALL-E? Will there be a time when a Hal decides that eliminating us may even be necessary to create scholarship?
I suspect that this fear is unfounded. Since the end of the last century, computer programs have been able to beat the best chess players in the world, but chess competitions continue to take place. And the inventions of the bicycle and the automobile did not end marathon races. In all likelihood, humans will remain essential to bioethics. The true value of bioethics (like most academic fields) is in the engagement with others, the conversations we have to raise new questions and find new answers.
In 20 years, what might my future self (or an AI) make of these speculations about AI and scholarship? If my 2026 reflections on my 2006 essay are any indication, speculations can be way off base or they can be prescient.
My suspicion about online journals – that they are categorically less authentic or scholarly than paper journals – hasn’t held up now that most journals are online. But my suspicion about articles being lost was prescient.
When I began to write this piece, I assumed that “File Not Found” was a relic. But it turned out that my first essay for Bioethics Forum was, in fact, not found there; it existed only on the Wayback Machine, a website that archives parts of the internet that are “lost.” My essay has since been restored to the Forum – by a human editor.
Tod Chambers, PhD, is an associate professor at the Northwestern University Feinberg School of Medicine and a Hastings Center Fellow.
