AI in healthcare

AI is showing up in hospitals, clinics, and health apps. This page explains what “AI in healthcare” means and where it shows up. This is educational only. It is not medical advice.

What “AI in healthcare” means

In healthcare, AI can mean models that predict or classify (like a risk score) and generative AI that drafts text. Some software can be regulated as software as a medical device.

Common uses (where it shows up)

• Clinical documentation help: draft notes and summaries for a clinician to review.
• Medical imaging support: flag possible findings in images for a specialist to double-check.
• Triage and routing: help route messages and decide who needs attention first.
• Drafting and summarizing text: draft a note with tools like Abridge or Suki (still needs clinician review).
• Simplifying patient instructions: rewrite explanations in plain language with tools like Grammarly or DeepL Write.
• Billing and coding support: suggest codes or extract fields (still needs human review).

Dive deeper with BonsAI Chat

What AI is good at (and bad at)

• Good at: spotting patterns in images and text. (NIST AI RMF)
• Good at: drafting summaries for a clinician to review.
• Bad at: making final clinical decisions. People must stay in charge.

Risks you must take seriously

• Made-up facts: the model can sound confident and be wrong. (NIST AI 600-1)
• Privacy: patient data must be protected.
• Harm: a bad summary or suggestion can affect care.

How to use AI safely (simple checklist)

• Do not paste patient identifiers unless your system is approved for it.
• Require clinician review before anything reaches a patient.
• Test on edge cases and monitor errors. (OWASP LLM Top 10)

How rules and regulators think about it (high level)

• Privacy and safety rules still apply (HIPAA, FDA, local rules). (OECD AI Principles)
• Some tools can be regulated as medical devices.

Questions to ask before you trust a tool

• What does the tool do when it is unsure? (NIST AI RMF)
• Can we see errors, logs, and evaluation results?
• How do we keep patient data safe?

Sources

• FDA: AI/ML Software as a Medical Device
• NIST: AI Risk Management Framework (AI RMF)
• NIST: AI 600-1 (Generative AI Profile)
• OWASP: Top 10 for Large Language Model Applications
• OECD: AI Principles