Skip to content
Roadmap for Healthcare AI

About this roadmap

A learning map for healthcare professionals: what to learn about AI, in what order, why it matters, and where to learn it.

Who built it

Built by Dr Karthik Deegutla, anaesthesiologist. karthikmd.com

An AI tool (Claude) helped find resources and write first drafts. Nothing was published on the AI's word alone: every resource went through the checks below, and each check is recorded.

How resources are checked

  • Blank beats wrong. Nothing is invented. When a detail such as a price, duration, CME credit or certificate could not be confirmed, it is shown as “Not listed” rather than guessed.
  • Two checks before anything is published: the link is opened and must lead to the stated resource on the provider's own site, and the resource must be clearly relevant to the track it sits in. Each card shows the date its link was checked.
  • A few major sites (for example some journal and NHS pages) block automated visits. Those resources are shown with a “Link not machine-checked” tag.
  • Journal papers are matched to their official DOI record (title, journal and year must match), and their open access status is looked up, so each paper shows whether a free full text exists.
  • Every link is re-checked weekly. Broken links are flagged for review and fixed or removed by hand, never automatically.
  • “Editorial selection” picks are suggestions with their criteria stated, not rankings. Topic explanations are plain-language summaries; feedback is welcome, and every one has a link to report an error.

How to use AI safely in clinical work

This roadmap never implies “AI says X, therefore the clinician should do X”. The safe pattern is:

AI output → verification → clinical judgement → appropriate action

Keep clear the difference between educational information, research assistance, decision support and clinical decision-making. Only the last is yours, and it stays yours.

Not clinical advice

Nothing on this site is clinical, legal or regulatory advice. It is an educational map of learning resources. Always follow your own clinical judgement, local policies and the primary sources, and check with your organisation's governance or regulatory lead before using any AI tool with patients or patient data.

Report a problem or suggest a resource

Use the report form to tell us about a broken link, a wrong or outdated detail, or a resource worth adding. Every card and explanation also has its own “Report a problem” link, which fills in what you are reporting about.

You can also email the editor.

Version v1.0. See the changelog for what has changed.