Medical emergency? Call 911. This site is education for professionals and the public, not medical advice.
DR GPT™ LabsIQDR GPT™ Intelligence
Briefing · November 12, 2026

Big Tech went all in on healthcare AI in 2026. Here is what it means.

By Harvey Castro, MD, MBA, board-certified emergency physician

The short version: in 2026, four of the largest technology companies each launched a healthcare AI product. Three of them are chasing the same unglamorous work: pulling records together, checking coverage and moving prior authorizations. That is good news for overloaded staff, and it raises the stakes on how hospitals review what these tools produce.

What each company shipped

CompanyProductMain buyerWhat it does
GoogleGemini for healthcare (U.S. preview, Nov 10, 2026)Health systems and health plansAgents that pull records, claims and coverage together for prior authorization, record review and member service [g1]
OpenAIOpenAI for Healthcare (Jan 8, 2026); ChatGPT for Clinicians (Apr 2026)Health systems, then individual cliniciansA HIPAA-ready ChatGPT workspace and API for organizations, plus a clinician product that turns tasks like referral letters and prior authorizations into reusable steps [o1] [o2]
AnthropicClaude for Healthcare (Jan 11, 2026)Providers, payers and life sciencesConnectors to the CMS coverage database, ICD-10 and the NPI registry, plus skills for prior authorization review and FHIR development [a1] [a2]
MicrosoftCopilot Health (announced Mar 12, 2026; U.S. preview May 29, 2026)Consumers and patientsOne place for a person's records, wearable data and lab results, with AI that explains them [m1] [m2]

What I see from the ER

1. The fight is over paperwork, not diagnosis. Google says Gemini for healthcare connects to more than 400 healthcare applications, including EHRs, and its launch customers include HCA Healthcare, Northwell Health and Optum [g1]. Anthropic built prior authorization review into Claude for Healthcare [a1]. I think that is the right place to start. Prior authorization and record hunting steal hours from clinicians, and a wrong draft there is easier to catch than a wrong diagnosis.

2. Read-only is a feature. The safest of these tools read the chart and draft, while a human signs. When a tool can write orders or send messages on its own, the review burden goes up sharply. Ask every vendor which one you are buying.

3. Patients are getting their own AI. Microsoft's Copilot Health brings a person's records, wearables and lab results into one place and says that data is not used to train its models [m1]. OpenAI has a consumer health product too [o1]. Expect patients to arrive with AI summaries of their own charts. Clinicians need a calm, consistent way to respond.

4. HIPAA readiness is now table stakes. OpenAI says content shared with ChatGPT for Healthcare is not used to train models, under a business associate agreement [o1]. That is necessary, and it is not enough. The questions that matter now are who reviews the output and what happens when it is wrong.

Five moves for hospital leaders this quarter

  1. Pick one administrative workflow, such as prior authorization, and measure its time and denial rate before any pilot.
  2. Require a named human reviewer for every AI draft that touches a chart, claim or patient message.
  3. Ask vendors to show results on patients like yours, across language, age and payer type.
  4. Write a one-page script for clinicians on how to respond when a patient brings an AI summary.
  5. Name one executive and one clinical lead accountable for AI outcomes, and review them every quarter.

For the full vendor checklist, see my 10 questions to ask before you buy healthcare AI.

Sources

I am a physician, not a lawyer or a financial advisor. This briefing is analysis and education, not legal, compliance or investment advice. Product names belong to their owners, and I am not affiliated with any company mentioned.

Written by , board-certified emergency physician. Last reviewed November 12, 2026.

About the author.
Want this briefing live for your board?
Harvey Castro, MD, MBA: board-certified emergency physician, 6x TEDx speaker and healthcare AI keynote speaker, in English and Spanish.
Book Harvey to speak
Get my free DR GPT™ newsletter
Healthcare AI in plain language from an ER physician, about twice a week.
Subscribe free