Artificial intelligence will transform healthcare.
That is no longer controversial.
What is still under-discussed is whether the assumptions being embedded into Health AI systems are still biologically valid after the pandemic.
This conversation exists because that question matters more than any pilot, demo, or procurement decision.
∗∗∗∗∗
“One of the best interviews I have ever done!”
Dr Robert Pearl
Dr Robert Pearl is one of the most experienced system-level voices in modern healthcare:
- Former CEO of Kaiser Permanente, responsible for care delivery to over 12 million people
- Professor at Stanford University School of Medicine and Business
- Author of ChatGPT, MD and Mistreated
- Regular contributor to Forbes
- Advisor on AI, health systems, and patient-centred reform
He understands AI not as novelty — but as infrastructure.
What this conversation actually explores
This is not a generic conversation about whether AI will change healthcare as that question is already settled. Instead, the discussion explores why narrow AI succeeded in constrained domains and how that success obscured deeper structural weaknesses in healthcare data. It examines how generative AI fundamentally changes what is possible in diagnosis, prevention, and care coordination, while simultaneously raising the stakes for data quality, clinical context, and outcome integrity. It also addresses why clinicians must actively shape this transition, rather than assume it will be benign if left to technical or commercial forces alone.
Crucially, the conversation surfaces a layer that is often missing from Health AI discourse: biological context. Much of the current debate focuses on compute, models, governance, and ethics. All of these matter. But AI systems also silently encode assumptions about physiology. Post-pandemic medicine is characterised by immune dysregulation, altered inflammatory baselines, delayed and non-linear disease expression, and changing risk profiles across age groups. If these shifts are not explicitly recognised, AI systems will optimise against a version of medicine that no longer exists.
The full conversation is presented below. It is intended for clinicians, health system leaders, AI developers working in healthcare, regulators, and policy designers who are responsible for deploying AI in real populations rather than controlled demonstrations. The aim is not persuasion, but clarity.
A note from Dr McMillan
I did not conduct this interview to “cover AI.” I did it because the Health AI transition is accelerating faster than our biological understanding is being updated. That mismatch, rather than AI itself, is where future harm will originate. This conversation is one contribution toward closing that gap before scale, not after.
If you are working on Health AI deployment, clinical AI governance, population-level modelling, or policy decisions tied to AI in healthcare, kindly recognise the importance of post-pandemic biological context.
You are welcome to continue the conversation.
First Published Paper on COVID-19 Autoimmunity
Involving Plasma/Serum ACE-2
