Healthcare AI Keynote Speaker: What Hospital Leaders Need to Hear in 2026

I have given AI keynotes to rooms full of hospital executives, and I can usually tell within ninety seconds whether the room is going to move.

The rooms that move are the ones where somebody on stage says the uncomfortable thing out loud. The rooms that do not are the ones where the speaker plays a montage of robots, quotes a market-size number, and leaves everyone entertained and unchanged.

If you are planning a healthcare conference, a leadership retreat, or a board education day for 2026, here is what I think your audience actually needs to hear, and what separates a keynote that produces action from one that produces polite applause.

Five messages hospital leaders need in 2026

One: the buying already happened, the governing did not

Adoption is no longer the story. Among US hospitals running Epic, 62.6 percent had adopted ambient AI documentation tools as of June 2025 (Yang F, Graetz I, American Journal of Managed Care, 2026).

Oversight is a different picture. In a Black Book Research survey of 182 hospital leaders in late 2025, only 29 percent had implemented and enforced policies covering AI model inventory, lineage, and sign-offs, and only 22 percent were highly confident they could produce a full AI audit trail within 30 days (Black Book Research release).

These studies measure different things in different samples. They can inform separate discussions about adoption and governance, but their percentages do not show how many hospitals adopted AI without adequate oversight.

Two: the vendor's accuracy number is not their accuracy number

The Epic Sepsis Model reached hundreds of US hospitals before independent evaluation. Researchers at Michigan Medicine studied it across 38,455 hospitalizations and found a hospitalization-level area under the curve of 0.63, with 67 percent of sepsis cases missed and 18 percent of hospitalizations hypothetically flagged. The analysis was retrospective, used a score threshold of at least 6 for these misses and hypothetical flags, and excluded hospitalizations eligible for actual alerts. According to PubMed, that is Wong A, Otles E, Donnelly JP, et al., JAMA Internal Medicine, 2021 (DOI).

This example lands with clinical and executive audiences at the same time, which is rare. Clinicians recognize the alert fatigue. Executives recognize the contract.

Three: the regulator now has a calendar

On June 1, 2026, The Joint Commission launched a voluntary Responsible Use of AI in Healthcare certification built on five domains: governance, data management, risk and bias reduction, monitoring and validation, and transparency with education and training (Joint Commission announcement). State-law requirements are separate and vary by jurisdiction. Georgia's SB 444, effective January 1, 2027, requires a qualifying natural person to conduct utilization review with a clinical peer participating before AI or software issues an adverse determination; the software cannot supersede that peer's judgment.

Audiences respond to dates. Vague warnings about "coming regulation" produce nodding. A specific effective date produces a meeting the following week.

Four: the real case is capacity, not accuracy

In March 2024, the AAMC projected a national shortage of up to 86,000 physicians by 2036 (AAMC March 21, 2024 report announcement). That projection alone does not establish which staffing options will work for a particular health system.

That reframes AI from a technology purchase to a workforce strategy, which is the frame most hospital boards are already prepared to fund.

Five: judgment is the job that does not get automated

I still work clinically. At 2 a.m. in an emergency department, the hard part is almost never retrieving information. It is deciding under uncertainty, with an incomplete history, a scared family, and a bed you do not have.

A keynote that ends by telling clinicians they are being replaced sends them back to work demoralized. A keynote that shows them exactly which part of their work is theirs alone sends them back willing to try the tool.

What makes a healthcare AI keynote land

Three things, in my experience.

A working clinician on stage. Rooms full of physicians and nurses have a fast, accurate detector for people who have never carried a pager. Credibility is not a bio slide. It is a story about a specific patient at a specific hour.

Real citations, on screen. Healthcare audiences read. Journal names, dates, and effective dates earn a level of attention that stock photography never will.

Something the audience can do Monday. A framework they can bring to their own committee. A short list of questions for their next vendor meeting. One decision they can make before the closing reception.

Signature talks

AI in Healthcare: The Good, the Bad, and the Unknown. An honest tour of what is working, what is quietly failing, and what nobody has solved. Best for mixed audiences of clinicians, IT, and executives.

The 2 a.m. Decision. Clinical judgment under uncertainty, and where AI helps and where it gets in the way. Built for physician and nursing audiences.

Credential the Model Like a Clinician. A governance framework that maps AI oversight onto medical staff credentialing, privileging, and peer review. Built for boards, quality leadership, and C-suites.

From Orbit to the ER. Lessons from building AI systems in aerospace applied to healthcare decision-making. Works well as an opening or closing keynote for innovation-themed programs.

Talks are customized after a planning call. I ask for the audience mix, the three outcomes you want, and any internal initiative the keynote should support.

Audience fit

Health system leadership retreats and board education sessions. Hospital association and society annual meetings. Health IT and digital health conferences. Physician group and specialty society meetings. Payer and life sciences leadership events. Government and ministry-level health policy convenings.

Sessions are available in English and Spanish. I am a native Spanish speaker and present regularly for Latin American and bilingual audiences, which is a thin field in healthcare AI.

Formats

Sixty-minute keynote with moderated question time. Forty-five-minute keynote for tight general sessions. Half-day board or executive workshop, built around your own AI inventory. Fireside chat or moderated panel. Virtual keynote for distributed leadership teams.

Why a physician for this topic

I am a board-certified emergency physician and a 5x TEDx speaker. I have written more than 45 books, many on healthcare AI and physician leadership, including AI in Emergency Medicine with Wiley and AI and Healthcare, 2nd Edition with Routledge. I served on Singapore's Ministry of Health Regulatory Advisory Panel through March 2026, and I advise the Texas Medical Association's Committee on Health Information Technology, which represents more than 55,000 Texas physicians. I taught in the UTSA and UT Health Medicine and AI dual-degree program, and I advise the CEO of Phantom Space. D Magazine named me ER Doctor of the Year from 2014 through 2022.

I mention all of that for one reason. When a hospital board asks whether a model should be allowed near a patient, the answer needs to come from someone who has stood at the bedside and someone who has read the regulation. Most speakers on this topic are one or the other.

Six questions to ask any healthcare AI speaker

Planners tell me they get pitched constantly and have no reliable way to sort the field. These six questions do most of the sorting.

1. Do you currently practice, or have you practiced clinically? Ask when they last saw a patient.
2. What sources will appear on your slides? A speaker who cannot name journals in advance will not name them on stage.
3. How will you customize this for our audience? Listen for questions back about your attendee mix.
4. What will attendees be able to do the Monday after? If the answer is inspiration, keep looking.
5. Have you spoken to a board specifically, and how was that different? Board audiences need governance and liability, not demos.
6. What do you say when someone asks whether AI will replace clinicians? The answer reveals whether the speaker understands the room.

If a speaker answers all six crisply, your program is probably in good hands, whether or not that speaker is me.

How to book

Tell me the date, the audience, and the outcome you want. I will tell you in one call whether I am the right speaker for your program, and if I am not, I will say so.

Book Harvey Castro, MD, DR GPT™ for your board, leadership retreat, or healthcare AI event.

www.HarveyCastroMD.com | /HarveyCastroMD | #DRGPT

Related DR GPT™ reading

• AI Governance in Healthcare: A Physician's Framework for Boards and Executives
• What Every Hospital CEO Should Know Before Buying AI

More from DR GPT™: who DR GPT™ is, healthcare AI keynotes, speaking, TEDx talks, books, and the media kit.

Book Harvey Castro, MD, MBA, known as DR GPT™, for your board, leadership retreat, or healthcare AI event. Start the conversation.