Imagine two healthcare futures: One driven by artificial intelligence with no human intervention and the other that’s entirely human, with no AI support.
Which would you choose?
That question kicked off a lively and often funny debate, one of many panels and presentations during Carey’s two-day 2026 Health Care Business Conference in early March.
The conference, themed Disruption to Discovery: Innovation in the Business of Health, capitalized on the 150th anniversary of Johns Hopkins University’s founding as the nation’s first research institution to drive the narrative about what’s coming next. It brought together powerhouses in academia, clinical practice, and health-related business to focus on the challenges and opportunities of AI in healthcare. Guests from across Johns Hopkins University as well as the Sylvester Comprehensive Cancer Center at the University of Miami, AstraZeneca, and Harvard Medical School, among others, discussed the many ways that new technologies are filling gaps in healthcare access and accelerating the pace of therapeutic discoveries.
Yet the road between even the best-intentioned idea and its successful execution is paved with hurdles, from investor buy-in to regulatory approvals to plain-old getting it right.
Do we have to choose?
The humans-versus-machines debate was posed in good fun, with interdisciplinary panelists recognizing that the all-or-nothing choice was unlikely in a world where artificial intelligence is already embedded in many aspects of healthcare. But the points debaters raised were worth serious consideration.
On the AI-only side were Bernard T. Ferrari Professor of Business Tinglong Dai and Computer Science PhD candidate Samuel Lefcourt. Their opponents on “Team Unplug” were Peetz Family Professor of Leadership Christopher G. Myers and Odia Kane, a postdoctoral fellow at the Johns Hopkins Institute for Assured Autonomy at the Whiting School of Engineering and the Berman Institute of Bioethics.
“Yes, AI can be terrible, but sometimes humans can be even worse,” said Dai. “AI is getting better. We are not. When we think humans can ensure safety, we have to think of the saying that ‘to err is human.’”
Myers countered, “To hallucinate absolute garbage nonsense is AI.”
He noted that AI often gets diagnostic challenges wrong, particularly in unusual cases.
“AI is essentially a tool of averages,” he said. “Averages are meaningless to the individual. We want individualized care.”
Perhaps nowhere would this be more evident than in precision oncology. Because each case is highly individualized, this evolving field of cancer care “generates enormous volumes of molecular and clinical data,” according to Dr. Chinmay Jani, chief fellow of Hematology and Oncology at the University of Miami’s Sylvester Comprehensive Cancer Center. The integration of those data analytics and AI into clinical workflows can be overwhelming to clinicians. Jani, along with his Sylvester colleague and director Dr. Stephen Nimer, Dr. William Nelson from the Sidney Kimmel Comprehensive Cancer Center at Johns Hopkins, Dr. David Ryan from Mass General Brigham Cancer Institute, and Dr. Thomas George from the University of Florida Health Cancer Institute, spent a session discussing the infrastructure, technology, and strategic investments required to support the care each patient needs as the science of oncology advances.
AI can’t understand communities
AI can make mistakes, but perhaps more importantly, people like being cared for by fellow human beings. And on the business side, human relationships are crucial to every stage of healthcare innovation, from bringing new funders aboard to earning the trust of patients and practitioners.
Access to care and sustainable models all over the world, with disparate regulations, languages, and levels of education and technological exposure, is a key element of global health, earning its own panel with experts from entrepreneurial, academic, and clinical fields. And there’s a critical need in the industry that poses a particular interest for AI. The panelists agreed that healthcare shortages create opportunities for AI to fill gaps, particularly in underserved areas. But that’s also a risk. Underserved areas need an understanding of community challenges. For that, Kane said, AI cannot be a substitute.
“Clinicians are members of a community, building relationships in addition to providing care,” she said.
In a session focused on the implications of using artificial intelligence to fill gaps in mental healthcare, School of Medicine Professor and Vice Dean for Clinical Investigation Gail Daumit also highlighted the equity issues that can accompany AI, with some populations having greater access to new technologies than others.
“We need to think about blended care models so these tools are connected to professionals when needed,” she said.
Elanah.AI CEO Josh Otero said his product, tailored to military veterans, provides mental health support without being a substitute for therapy. Key to that tailoring is the understanding that the veteran population is far more likely to die by suicide than non-veterans. That understanding is crucial to building an AI agent that can offer veterans support at 2 a.m.
“This [technology] is built for the community, but it’s also built with the community,” he said.
Ananya Joshi rounded out the panel. Joshi is an assistant professor of Psychiatry at the Johns Hopkins School of Medicine and also an assistant professor in the Computer Science department at the Johns Hopkins Whiting School of Engineering. She leads the PALS Lab, which focuses on psychiatry-assistive learning systems. Top-of-mind for all three panelists were tragic news stories of people who turn to AI for companionship or mental health support, only to be harmed by the very algorithms they sought out for help. They said it’s important that AI tools use data and deliver results that are safe and effective for their intended populations.
“We should use AI to augment our human capabilities,” Joshi said. “That is really what we want to work toward.”
Where human-AI collaboration works
One example of AI and clinicians working together is diabetic retinopathy screening, which has improved outcomes and, not coincidentally, been more convenient for patients and clinicians. School of Medicine Associate Professor of Pediatrics Risa Wolf directs the school’s Pediatric Diabetes program, which she brought to Johns Hopkins in 2018, shortly after it won Food and Drug Administration approval.
“We needed a lot of implementation and validation, but it is achieving the goal of getting kids to get the screening,” she said.
In a panel on AI’s role in propelling drug discovery, Guardant Health Chief Medical Officer Craig Eagle noted that AI’s potential, particularly to accelerate discovery and improve treatment decision-making, is enormous.
“Every day I wake up and hear about something I didn’t realize it could do,” he said.
And yet, it’s still a lift to “bring the people along,” he said, requiring transparency and ensuring that it both delivers value and protects patients. That theme echoed across multiple panels, whether about specific developments or general applications, whether from corporate industry experts or researchers. One in particular was the panel focused on women leading innovation in health, organized as a nod to Women’s History Month.
“Innovation really sticks when you have buy-in from the beginning, when you include broad representation and explain the ‘why,’ and you’re not thrusting it on people after it’s fully baked and saying, ‘here, adopt it,’” said Neelima Denduluri, vice president of medical affairs at AstraZeneca.
“There are challenges to getting it to the people who need it,” said Carey Professor Stacey Lee. “It has to be a multi-person effort.”
Bringing the conference together
Carey’s Health Care Business Conference was organized by staff, faculty, and students, many of whom brought in the guests who served on panels. Lead student organizer Shreyans Singhvi, MD is pursuing a dual MBA and master’s degree in Health Care Management at Carey. He said his goal for the conference was to bring together experts in business, medicine, ethics and technology to showcase Johns Hopkins leadership and provide networking opportunities. About 200 people attended the two-day event, he said, with more than 150 dialing in for remote participation.
this year, in addition to the AI debate, was a session hosted by the Johns Hopkins Institute for Planetary Health, a year-old research group that spans Johns Hopkins University, to generate ideas for safeguarding health on a rapidly changing planet.
The 2026 Health Care Business Conference was sponsored by Guardant Health. Additional funding was provided by Abbott, Gilead Sciences, Kaiser Permanente and the Hopkins Business of Health Initiative.
