If AI Can Prescribe, Who's Liable When It Gets It Wrong? Four Experts Weigh In

Key Podcast Moments

• The Prescriber Is the Last to Know
  Simon Chang explains why a prescription travels one way through the system: the pharmacist learns a drug is expensive or out of stock first, the patient second, and the physician often never finds out at all.
• AI Prescribing Without a Doctor in the Loop Simon reveals that some companies are already sending AI-generated prescriptions with no clinician involved, and that pharmacies have no way to tell because the prescription standard cannot label AI involvement.
• One Flawed System, Many Patients David Simon explains why AI changes the liability math: instead of isolated errors spread across individual clinicians, one flawed system can repeat the same mistake across many patients.
• It's Not Bad Decisions. It's Broken Handoffs Deepika Srivastava shares what a malpractice carrier sees across thousands of closed claims: communication failures between steps in care often do more harm than a bad clinical call.
• Doctor Picks the Drug, Patient Picks the Pharmacy Tanvi Patel and Simon describe a future where the digital prescription goes back into the patient's hands, with price, stock, and delivery speed visible, while the physician stays focused on clinical decisions.

Prescribing in 2026 still runs on one-way messages, phone tag, and guesswork, even as AI writes our notes and reads our imaging. In the first How I Doctor Roundtable, Dr. Graham Walker brings together four people who each see the problem from a different seat: Dr. Simon Chang, an internal medicine physician, AI Task Group lead for NCPDP, and founder of eNavvi; Tanvi Patel, Vice President and General Manager of Amazon Pharmacy; Deepika Srivastava, Chief Operating Officer of The Doctors Company; and David Simon, Associate Professor of Law at Northeastern University. Two questions drive the hour: why is prescribing still this hard, and what happens when AI starts making more of these decisions for us?

Simon explains that after a physician signs a prescription, it travels one way to the pharmacy, so the pharmacist and patient usually learn about price or stock problems before the prescriber does. He also flags that some companies are already sending AI-generated prescriptions with no doctor involved, and pharmacies can't tell.

Tanvi looks at the same problem from the patient's side, arguing that e-prescribing quietly shifted the job of choosing a pharmacy onto the physician. Her fix is to put the digital prescription back in the patient's hands, with price, stock, and delivery speed visible up front.

David says tort law already handles a physician's prescribing mistakes well, but AI changes the scale: one flawed system can repeat the same error across many patients. His advice is to tell patients when AI is in use and to vet any tool's error rates before relying on it.

Deepika explains that a malpractice carrier doesn't start by assigning blame, and that across thousands of claims, broken handoffs and poor communication cause more harm than bad clinical decisions. She warns that once AI starts choosing for the physician, the risk changes, so guardrails and good governance matter and physician decisions should never be silently overridden.

The conversation closes with a lightning round on what good looks like two years from now. Simon and Tanvi both describe a world where physicians decide the clinical question and patients keep informed control of the logistics. David hopes for a coherent federal strategy that divides labor sensibly between federal and state regulators, and Deepika wants laws and rules that help physicians and patients participate in decisions rather than be bypassed by them. Different seats, same conclusion: AI can take friction out of prescribing, but physicians have to stay accountable for, and in charge of, the decision.

This episode is presented in partnership with eNavvi, a clinician-built digital prescribing platform.

ePrescribe anytime anywhere with eNavvi

Top 4 Takeaways

AI prescribing is no longer hypothetical, and the people who build it, sell through it, insure it, and litigate it agree on one thing: the rules are still being written. Here are the four takeaways physicians should walk away with.