Ethics at the Edge: Bioethics, AI, and the Courage to Lead

This transcript of the discussion has been edited for clarity and length.

Ethical decision-making in healthcare is not just a theoretical exercise; it is a practical and urgent responsibility that requires courage, transparency, and active engagement. Physician leaders must build robust ethical infrastructures, stay prepared for complex dilemmas, and embrace their duty to combat misinformation and inequity.

From Polio Ward to Bioethics Pioneer

Mike Sacopulos: Professor Caplan, you’ve been at the forefront of bioethics for decades. Can you describe your career trajectory and what initially drew you to the field?

Arthur Caplan, PhD: I’m one of the last people in America — if not the world — to have had polio. I contracted it at age six and was hospitalized at Mass General, probably around 1956 or 1957, just missing the polio vaccine by months. Even as a young child, the experience got me thinking about how patients were treated. Visiting hours were severely restricted. Children who were very ill disappeared from the ward, and we were told they had gone home. We knew they hadn’t. We knew they were dying.

That experience — the truth-telling, the isolation from family, the boredom of long-term illness — oriented me toward questions about how we care for people in institutions. Later, I went through years of rehabilitation and encountered a different kind of ethics question: the expectation that patients, even young ones, should be “motivated” and “compliant.” I was ten. I wanted to watch television.

I went to Brandeis for college during a period of intense public debate — civil rights, Vietnam, women’s rights. I got caught up in those arguments. I tried medical school at Columbia, then shifted to a graduate philosophy program, although they made it clear they weren’t particularly interested in the ethics of medicine. I earned my PhD anyway, writing a dissertation on the philosophy of science and creationism, and eventually returned to Columbia to teach medical ethics. My first course was a disaster — I taught it like a philosophy seminar, and by the end I had six students left, none with English as a first language. The dean told me I was doing it wrong. In medicine, he said, you teach by cases. I switched, and the course took off. That was the beginning.

Bioethics vs. Philosophy: Two Ways of Looking at the Lake

Sacopulos: It seems like bioethics and academic philosophy attract different kinds of students, even when the subject matter overlaps. Is that your experience?

Caplan: Completely. A lot of PhD philosophy students are drawn to abstract, idealized reasoning — what would perfectly rational, disembodied beings agree to if they had no self-interest? That’s a powerful technique, but it’s designed to simplify reality in order to understand it. Think of a biologist who studies a lake by building a mathematical model rather than wading in and sampling the sediment. Clean, but removed from the actual complexity.

Bioethics starts in the lake. It wants to know all the variables — the messiness, the competing pressures, the real people involved. And if someone objects that we’ll never arrive at a perfect ultimate principle, the bioethicist’s response is: we don’t need one. We just need enough consensus to solve today’s problem. If people broadly agree that patients have the right to consent to or refuse treatment, that’s sufficient. Our work is done.

Building an Ethics Infrastructure That Actually Works

Sacopulos: My impression is that ethics committees are underused in many health systems. What does an effective ethics infrastructure look like, and what role should physician leaders play?

Caplan: Start with the basics. Every hospital and every nursing home or long-term care facility should have an ethics committee that meets regularly, keeps minutes, and maintains active liaison with the institution’s legal counsel. If the committee doesn’t know what the law requires, it needs a fast path to finding out.

Beyond the committee, you need an ethics community. That means ethics rounds, regular education for nurses, social workers, patient advocates — everyone who touches care delivery. Policies need to exist on foreseeable dilemmas — parental refusal of treatment for a child, for example — and people need to actually know what those policies say.

I also think leaders should periodically hold a fire drill. Construct a scenario, call it in, and see what happens. Who gets involved? What do people do? What’s the public position of the institution going to be? You want to test your systems before you need them. Ask your hospital attorney for the two or three scenarios that scare them most — you’ll have a simulation in about 24 hours.

The Bioethical Frontiers of the Next Decade

Sacopulos: What do you see as the defining bioethical challenges of the coming decade?

Caplan: Four areas stand out.

The most obvious is AI. It’s already deeply embedded in healthcare, and it raises questions that legislation and regulation have not yet addressed. When must you tell a patient that AI was involved in their care? When is an AI system competent enough to take on a particular task? Who bears liability when AI contributes to a medical error? And what are we doing to protect the privacy and confidentiality of health data in an environment where regulations like HIPAA were written before the Internet existed? These are not theoretical questions. Leaders need to be working with their professional associations to set standards now.

The second area is neuroscience. The study of the brain is beginning to catch up to genetics in its clinical and ethical implications. Brain imaging, implants, virtual reality interventions, pharmacological monitoring via scanning — all of these raise questions about who owns the data, who can make a diagnosis based on a scan, and what happens when a device that was supposed to restore vision or reduce anxiety doesn’t deliver. Research to reach these interventions raises its own set of issues.

Third is cost and rationing. We still have a healthcare system that cannot control its prices. Gene therapies costing $2 million per patient are not hypothetical — they exist now. Will access be limited to those who can pay? Will health systems develop fair, evidence-based processes for deciding what goes in the formulary and what doesn’t? Leaders must grapple with making those processes transparent and equitable, not driven by the preferences of the most productive surgeon on staff.

Fourth is research ethics. Our current framework — informed consent, IRB review, conflict-of-interest management — was built in response to historical abuses: the Nazi experiments, Tuskegee, Willowbrook. It was designed to protect subjects. But today’s patients — especially those with serious illness — are less concerned with protection than with access. They want to get into trials. They want to move faster. I predict we’ll see a meaningful shift toward faster approvals with more real-world post-approval monitoring, and leaders need to be prepared to manage both the cost implications and the ethical ones.

Compassionate Use: A Model for Managing the Impossible Request

Sacopulos: You co-chair the Compassionate Use Advisory Committee for Johnson & Johnson’s Janssen Pharmaceuticals. What should physician leaders know about navigating requests for unproven therapies?

Caplan: First, every institution needs a policy — developed with IRB and legal input — about how it will handle requests for unapproved drugs, devices, or vaccines. We’re already seeing early-stage cancer vaccines that show real promise but haven’t completed trials. Requests for these are coming.

When these requests arrive, the question isn’t just whether to honor them. It’s whether you’ve built the infrastructure to evaluate them fairly. Any request should come through a physician, not directly from a patient. It should be reviewed by the human experimentation committee, with legal involved to address liability and payment questions, and PR in the loop, because there is no worse headline than a patient dying after receiving an unproven treatment at your institution.

The compassionate use advisory committee model I helped develop was initially oriented toward industry — companies like Johnson & Johnson were receiving direct requests from physicians at hospitals around the world and weren’t managing them well. We created an independent international committee to review those requests against principles of fairness, evidence, and transparency, with patient advocates involved. The result has been access for many more patients than would have participated in clinical trials alone.

For hospital leaders, the practical lesson is this: know the landscape. If a request comes in, who handles it? Where does it go? Does the company involved have a compassionate use process? Can you liaise with it? Run a fire drill on this scenario. The calls come from prominent people — the senator’s spouse, the local business leader whose child is ill. You need to be ready before the phone rings.

Misinformation, Courage, and the Duty to Speak

Sacopulos: Health misinformation — including from political figures — is actively shaping patients’ beliefs. How should physician leaders rethink their approach to informed consent in that environment?

Caplan: I’ll say something controversial: I don’t think leaders are exercising enough courage. Too much misinformation goes unrebutted. A lot of institutions are in duck-and-cover mode — worried about provoking state officials, losing research grants, triggering audits. I understand those pressures. But the long-term cost of silence is the erosion of trust in mainstream medicine. If patients can’t trust institutions, they don’t go to institutions. They go to wellness clinics, buy supplements from influencers, or pursue care through AI-generated recommendations delivered in a brown paper wrapper. Healthcare in this country runs on the lubricant of trust. Allow misinformation to fester, and that lubricant dries up.

What I advocate for is proactive community engagement — getting voices out into high schools, civic organizations, churches, synagogues. Identify faculty, physicians, and nurses who want to talk publicly and give them the support and the incentives to do it. Reward community education the same way you reward procedural efficiency. It doesn’t require the CEO to personally rebut every false claim. But when someone inside the institution wants to correct the record, they should get institutional backing, not a quiet request to stand down.

I’ll also note that misinformation isn’t only an external problem. I cannot find a single cancer center in the United States that doesn’t offer Reiki, aromatherapy, or other interventions with little or no evidence base. We tell students and trainees that medicine is evidence-based, and then we put these services on the menu. If we’re going to offer them, we should be honest about what they are and what they aren’t — and make sure we’re framing them accurately to patients.

Principles for Rationing Decisions

Sacopulos: From organ allocation to ICU beds during COVID, rationing decisions expose deep inequities in our system. What principles should guide physician leaders making those calls?

Caplan: Three principles, in brief:

Equal quality of care for all, regardless of socioeconomic status, background, or insurance coverage. We can’t always fully achieve it, but it must be what we aspire to. At NYU, we run both a premium facility and Bellevue, where the same physicians and the same standards apply, even without private rooms or gourmet meals.

Evidence-based decision-making. Maximize the allocation of interventions that have the strongest track record of efficacy. That doesn’t mean everyone gets the newest drug — it means making sure the proven therapies reach as many people as possible.

Dignity and respect for every patient, at every hour, from every member of the workforce — not just the day shift, not just clinical staff, but temporary workers, evening staff, weekend staff. Patients don’t distinguish between the experienced nurse and the temp. If their experience at 2 AM is poor, it shapes their perception of the institution. Leadership must model this and communicate it consistently.

The Future of Bioethics Education

Sacopulos: How is bioethics being integrated into medical training today, and what does the next generation of physician leaders need?

Caplan: The honest answer is that it was strong until recently and is now under pressure. Bioethics programs at many medical schools were subsidized by research overhead — indirect costs from grants that flowed into institutional budgets and, in part, supported ethics programs. As grant funding contracts and overhead rates fall, that quiet subsidy disappears. Bioethics is now having to justify itself on its own financial terms, and many programs can’t do that. There simply isn’t enough grant funding or philanthropy directed at the field.

That said, there are reasons for optimism. The most important site of bioethics training is not the medical school classroom — it’s residency and fellowship. That’s where young physicians are forming their values in real clinical situations, encountering patients who aren’t cooperative, families who aren’t grateful, requests that can’t be honored. At NYU, we are actively expanding ethics education into those settings.

Continuing education is also thriving. Online courses in transplant ethics, genetics and ethics, vaccine ethics — and, soon, more leadership-oriented ethics programming — are reaching practicing physicians and administrators who need it most. The combination of real-world practitioners and real cases is a rich environment for developing the next generation of physician leaders.

The field has come a long way from its early days, when a colleague and I used to joke that bioethics might last a year or two before it ran out of things to say. What I tried to do, and would encourage leaders to support, is keep it close to the public — not locked in academic journals, but present in communities, in the news, and in the conversations that shape how people understand medicine and trust the institutions that deliver it.

Listen to this episode of SoundPractice .