Anthropic has 22 free courses and 119 short tutorials online, including a 25-minute lesson that connects Claude directly to PubMed. Not one medical society has told its members it exists.
A colleague my age told me in June that he would start using AI seriously as soon as someone taught him properly.
He is a good doctor.
He was not making an excuse. He was waiting for a course to arrive from a society, a department, or a journal. Nothing is coming. The teaching already exists, it is free, and the part that matters to an ObGyn takes about six hours.
The site is called Claude Academy.
It is run by Anthropic, the company that builds Claude. Everything on it is free, and a free Claude account is enough to take the courses. The catalog sits on one page and holds 22 courses, 119 short tutorials, 148 worked examples, and 68 recorded talks. A course is a set of short video lessons with a quiz and a certificate at the end. A tutorial is a five to twenty-five minute lesson on a single task. There is no fee, no application, no institutional login.
Most of it is not for us. Three of the courses teach programmers to build software on top of Claude, and they run eight to nine hours each. Skip those. Here is the order I would give a colleague.
Start with AI Capabilities and Limitations. Thirteen lessons, about three and a half hours. It explains what the machine is actually doing when it answers you: predicting the next piece of text, one piece at a time, from patterns it learned. It explains what it holds in working memory and what quietly falls out. This is the course that tells you why a model can produce a reference that looks perfect and does not exist. A doctor who skips it will either trust the output too much or dismiss the whole field. Both mistakes come from the same gap.
Then Claude 101. Thirteen lessons, two and a half hours. Ordinary work: how to ask well, how to organize a project, how to connect outside tools.
Then AI Fluency: Framework and Foundations. Fourteen lessons, four hours. It teaches four habits, called delegation, description, discernment, and diligence. The last two are the ones medicine needs. Discernment is judging what came back. Diligence is owning it.
After that, one 25-minute tutorial is worth more to a working ObGyn than any of the courses: connecting Claude to PubMed. That is the National Library of Medicine database, searched and summarized inside the conversation, against real records instead of the model’s memory. A companion tutorial does the same for ClinicalTrials.gov, in 15 minutes, across a registry of more than 500,000 trials. If you write, review, or teach, do these two first and take the courses afterward.
Two more, depending on the work. If your day is documents, protocols, and files rather than chat, take Introduction to Claude Cowork, fourteen lessons, two and a half hours. If you want to build your own calculator or risk tool, as I do, Claude Code 101 is twelve lessons and one hour. If you run a residency or a course, Teaching AI Fluency was written for faculty who have to teach this to other people: seven lessons, four and a half hours. There is also a recorded session from April 2026 in which practicing physicians demonstrate tools they built themselves, with the safety and verification questions asked out loud.
Now the part the catalog does not tell you.
This curriculum was written by the company selling the product. That does not make it wrong. The lessons on capability and limitation are more honest than most of what medical societies have published about AI. But it is an owner’s manual, not an independent evaluation. It teaches you to use the tool well. It does not teach you to practice medicine with it.
No lesson tells you to check every citation against PubMed before it enters a manuscript. No lesson tells you what to do when the answer gives a confident and wrong magnesium dose in a woman with severe preeclampsia. Nothing in the catalog addresses whether a patient should be told that an AI tool helped shape her care, or whether she may decline it. That paragraph belongs in informed consent, and nobody has written it yet. Anthropic will not write it for us.
Conclusion
I offered to teach this material to two of our professional societies. Both declined. So the training gap in obstetrics is now a choice, not a shortage. Take three things: AI Capabilities and Limitations, Claude 101, and the PubMed tutorial. Six hours, three evenings, no cost. After that, “nobody trained me” is retired as a defense. What no course will hand you is the discipline that makes any of it safe: verify every number, every citation, every dose, before it reaches a patient or a page. The machine carries no license. You do.
Send this to the colleague who keeps saying he will get to AI eventually. Then take lesson one tonight. The series you were going to watch will still be there tomorrow.
ObGyn Intelligence is free because the work matters. A paid subscription keeps it independent.
Links
The full catalog, all 357 items on one page: https://academy.claude.com/all
AI Capabilities and Limitations (13 lessons, 3.5 hours): https://academy.claude.com/courses/ai-capabilities-and-limitations
Claude 101 (13 lessons, 2.5 hours): https://academy.claude.com/courses/claude-101
AI Fluency: Framework and Foundations (14 lessons, 4 hours): https://academy.claude.com/courses/ai-fluency-framework-foundations
Using the PubMed Connector in Claude (25 minutes): https://academy.claude.com/tutorials/using-the-pubmed-connector-in-claude
Using the ClinicalTrials.gov Connector in Claude (15 minutes): https://academy.claude.com/tutorials/using-the-clinicaltrials-gov-connector-in-claude
Introduction to Claude Cowork (14 lessons, 2.5 hours): https://academy.claude.com/courses/introduction-to-claude-cowork
Claude Code 101 (12 lessons, 1 hour): https://academy.claude.com/courses/claude-code-101
Claude Code in Action (9 lessons, 1 hour): https://academy.claude.com/courses/claude-code-in-action
Introduction to agent skills (6 lessons, 1 hour): https://academy.claude.com/courses/introduction-to-agent-skills
Teaching AI Fluency, for faculty (7 lessons, 4.5 hours): https://academy.claude.com/courses/teaching-ai-fluency
Claude Code for Healthcare: How Physicians Build with AI, recorded 23 April 2026: https://www.anthropic.com/webinars/claude-code-in-healthcare-how-physicians-are-building-with-claude
All recorded webinars: https://academy.claude.com/webinars