Healthcare AI isn’t just hype. In an industry where seconds count and complexity rules, artificial intelligence offers both a lifeline and a landmine for American medicine. As hospitals confront workforce shortages, patient volumes, and regulatory churn, AI’s potential to reshape decision-making, improve outcomes, and extend reach has never been more consequential.
But AI’s value isn’t theoretical. This is the era where an ER physician can, within days, publish a book mapping out how Apple’s Vision Pro and generative AI might transform emergency care, triage, and patient engagement. That physician is Harvey Castro, MD, MBA, an ER physician, entrepreneur, and AI futurist with deep operational and clinical experience. In his recent conversation on the American Journal of Healthcare Strategy podcast, Dr. Castro—now at {{Guest Workplace Name}}—pulled back the curtain on how AI is colliding with medical practice, why he’s bullish on next-gen tech, and what executives need to know to stay ahead of the curve.
If you’re navigating health system transformation, evaluating investments, or leading teams through digital disruption, his perspective is required reading. Here’s what you need to know.
Dr. Harvey Castro is an ER physician, entrepreneur, and self-described “Nutty Professor” of healthcare technology. He’s launched companies, authored books, and held C-suite roles—“I created about eight emergency rooms here in Texas and at some point had 350 employees,” he shares, illustrating his firsthand understanding of both the front lines and the boardroom.
What makes Castro credible to speak on healthcare AI?
Track Record of Innovation: From building a vitamin company and launching over 30 medical apps—including the first IV medications app for iPhone—to writing the first book on ChatGPT in healthcare in 2022, Castro moves fast and builds at scale.
Clinical Gravitas: Board-certified and active in emergency medicine, he grounds his tech vision in real-world clinical needs.
Futurist Lens: “I take pride in being a thought leader and a futurist,” Castro says. His work straddles the now and the next.
“I literally put on my doctor hat, my CEO hat, and started writing a book about different ideas—nothing that cannot be done today. Honestly, some of the crazy ideas I came up with are just taking different technologies and putting them all into one. I really think that is going to be the future.”
AI, paired with Apple Vision Pro, could revolutionize real-time care in the ER—if health systems act quickly.
Dr. Castro offers a concrete vision: “Imagine you’re my patient, and I have the Apple Vision Pro on. Seconds count in the ER. I might need your X-rays, your labs, your entire medical record. What if I could see all that, in real time, in my field of view, while taking care of you?”
Real-Time Clinical Data Access
“Being able to see your labs, X-rays, and chart while caring for you—maybe you’re unconscious or unable to answer—would be a game-changer.”
Language Translation at the Bedside
“What if the Vision Pro could translate in real time as I speak, giving you information in your language and me in mine?”
AI-Powered Chart Summarization
“Imagine having AI summarize your volumes of medical records on the fly, letting me query specifics while I’m with you.”
Efficiency: Time-to-treatment drops, cognitive load lightens.
Equity: Language barriers and data overload decrease.
Safety: The right information, at the right time, for the right patient.
But Castro is clear-eyed about the politics: “Departments that bring in a lot of profit—like neurosurgery—are able to get this technology first. Less profitable departments may lag, unless we can prove these tools save lives or drive efficiency.”
Widespread AI adoption in U.S. hospitals faces both economic and cultural headwinds.
Castro doesn’t mince words: “There’s always politics. Departments that are financially strong can say, ‘I want this new gadget,’ and get it. Others—often the ones losing money—are less likely to get this technology.” This resource gap risks exacerbating existing healthcare inequities unless leaders focus on measurable outcomes and scalable use cases.
Budget disparities between specialties and departments
Culture and “old ways” of working (yes, many hospitals still use fax machines)
Uncertainty about ROI, liability, and workflow impact
Fear of the unknown—from bedside nurses to C-suite skeptics
Despite these challenges, Castro’s optimism is grounded in recent momentum: “As we start creating use cases and showing that this is saving lives or becoming more efficient, hospital systems will end up funding it—even in departments that wouldn’t normally get this technology.”
Necessity—and frustration—drove Castro’s AI journey.
“I was coding a patient, gave the nurse an order, and she had to look up the dosage in a textbook, do the math, and it took forever. I thought, ‘There has to be a better way.’ So I taught myself to program and created the first IV app in the world.”
Fast forward to ChatGPT’s launch, and Castro recognized a similarly seismic shift: “It was my iPhone moment all over again. I thought, ‘This is going to change medicine.’” That realization fueled his book, ChatGPT and Healthcare, and catalyzed his mission to educate clinicians on AI’s possibilities and pitfalls.
Innovators are made, not born: Frustration + curiosity + action = transformation.
Tech adoption is a mindset before it’s a process: “I tend to be a very creative person in that space,” Castro notes.
Education is a force multiplier: Early adopters must become teachers, demystifying AI for others.
The biggest risk isn’t that AI will replace doctors, but that it will mislead patients or clinicians.
Castro doesn’t sugarcoat the hazards: “Non-physicians may use ChatGPT like Dr. Google. They’ll put in symptoms and, if ChatGPT says they’re okay, they may skip seeing a doctor. Outside the U.S., where you can get drugs at a pharmacy, this is even riskier.”
He’s wary of self-diagnosis, medication misuse, and “hallucinations”—AI-generated errors—which, even at a 4% error rate, are unacceptable when stakes are life and death. “Even if the newer models are saying they can get the error rate down to 4%, to me that’s 4% too much because this is life or death.”
Patients: “Use ChatGPT, but always run its advice by your doctor.”
Clinicians: “Physicians five years out of residency, with real clinical experience, are best equipped to discern when AI gets it wrong.”
Hospitals: Post clear warnings, educate both staff and patients on what AI can and cannot do.
Certain uses of AI can save time, but must be clinically verified.
Castro walks a fine line: “It’s tough because of medical-legal risk. Even medical students can be fooled by AI. The sweet spot is clinicians with enough experience to know when AI is wrong.”
Examples of “safe” AI support:
Personalized diet or wellness plans, reviewed by a doctor
Automated medical record queries and summarization—if the output is validated
Integrated tools (e.g., AI meal plans plugged into grocery delivery apps, but confirmed by clinicians)
“If it’s used together—doctor and patient working with AI—we’ll get great outputs. But it has to be verified.”
Data privacy laws like HIPAA and GDPR set a baseline, but “AI food labels” could bridge the trust gap.
Castro, an avowed capitalist, hesitates to advocate new regulation, but sees the need for transparency:
“If I had to pick, I’d create ‘food labels’ for AI. Anyone—doctor or patient—could pick up that AI and see, here are the positives, here are the negatives, here’s the training, here’s why it has bias, here’s how it should be used.”
Why not more? He sees current privacy protections as sufficient—“HIPAA already covers that.”—and worries about stifling innovation. But consumer-facing “labels” can inform choices without slowing progress.
AI’s “black box” reputation and uneven tech literacy fuel fear. The remedy: education and transparency.
Castro’s approach is hands-on: “I created my own AI Healthcare course, I’m giving lectures around the world. If we explain how this works, it takes away fear.” He suggests “knowledge graphs” tracing AI’s logic, so clinicians can see how conclusions are reached.
Key drivers of fear:
Opaque algorithms (the “black box” problem)
Variability in technology comfort across clinicians
High stakes—errors aren’t abstract, they’re personal and potentially catastrophic
“When there’s an outcome, create a knowledge graph so doctors can see how you got from point A to B to C to D.” Custom, clinically trained models for high-impact verticals (think diabetes, radiology) could further mitigate risk.
Find a pain point, combine it with passion, and don’t be afraid to do things differently.
Castro’s entrepreneurial arc is a lesson in problem-driven innovation: “I was told I was spending too much time with my patients, ordering too many tests. That pain point upset me, so I built a different culture: more time with patients, more praise for doing the right thing.”
He distills his philosophy into three steps:
Find your pain point—what frustrates or limits you in your current environment?
Get passionate—does the problem matter enough to fuel real change?
Think outside the box—“Just because something’s never been done doesn’t mean it’s wrong.”
Castro’s advice to aspiring entrepreneurs, especially in healthcare, is as much about mindset as it is about market opportunity.
Healthcare AI is already reshaping the clinical and operational landscape—if you know where to look, and if you’re bold enough to bridge the old and the new. Dr. Harvey Castro, MD, MBA, ER Physician & AI Futurist at {{Guest Workplace Name}}, exemplifies the blend of frontline experience, creative problem-solving, and operational rigor required to lead in this environment.
For U.S. healthcare executives, clinicians, and policy leaders, the call to action is clear:
Invest in educating your workforce—both to adopt and to question AI.
Demand transparency from technology partners (“AI food labels”).
Start piloting use cases that address real pain points—especially where time, access, and equity are at stake.
Maintain clinical oversight at every stage, never letting tech replace expertise.
“We need to be advocates for our patients… if AI is used together with doctors and patients, we’ll have great outcomes.”
Ready to lead the next chapter? The future is already in the ER—and in your boardroom.
