AI in health care: what works and how to start
How many are using it
of US health care and social assistance businesses used AI in the fortnight 21 Sep 2026 to 4 Oct 2026, by the US Census Bureau’s Business Trends and Outlook Survey (published 22 Oct 2026).
Businesses, not workers: a firm counts if it used AI at all. Every sector · Over time
What works here
All rated use cases →No clear effect so far: the studies do not show that AI scribes reduce clinicians' time on notes.
8 studies counted →Health careDo AI-drafted replies save clinicians time on patient messages?Not enough evidence yet to say whether AI-drafted replies save clinicians time on patient messages.
5 studies counted →Health careDoes an AI chatbot help doctors diagnose and plan treatment?Not enough evidence yet to say whether an AI chatbot helps doctors diagnose and plan treatment.
5 studies counted →What hasn't worked
Across all sectors →- AI drafts of patient replies, Tai-Seale et al. (2024): +21.8% Time spent reading a patient message (includes reading the AI draft).
- AI drafts of patient replies, Jeanselme et al. (2026): +8.6% Message turnaround time (patient submission to first clinician reply) after a clinician adopts the tool.
- AI drafts of patient replies, Jeanselme et al. (2026): Messages received after hours: Tool use linked to 4.52 percentage points more after-hours replies
- AI drafts of patient replies, Chen et al. (2024): Unedited AI drafts: 7.1% were judged to risk severe harm if sent as written.
- AI drafts of patient replies, Mandal et al. (2025): +7.3% Median time the message was open, with versus without the draft.
- AI help with diagnosis and treatment, Goh et al. (2025): +119.3 seconds per case Time per case.
- AI help with diagnosis and treatment, Everett et al. (2026): AI-as-second-opinion cases: Actionable-decision score fell after seeing the AI in 12 of 147 cases (rose in 52, unchanged in 83)
Rules to check first
In the US, software that diagnoses or treats is a medical device and needs FDA authorisation. 1,433 AI-enabled devices are on the FDA's list. FDA AI devices data →
Health privacy law (HIPAA in the US, GDPR in Europe and similar laws elsewhere) covers patient data you put into an AI tool. Check the vendor's contract before any patient data goes in.
Ambient scribes record the visit. In the trials, patients were told and agreed first; recording law differs by country and state.
AI that is, or is a safety part of, a medical device needing an independent conformity check counts as high-risk under the EU AI Act, with extra duties for makers and users. What applies when →
Before you start
Will AI help with your work? →- Start with paperwork, not diagnosisMost trials so far are about documentation, notes and messages, where a clinician checks every word before it counts; even there, results are mixed.
- Run a real comparisonMeasure time in the record system and burnout with validated surveys, for users and a comparison group, before and after.
- Keep the clinician responsibleEvery drafted note or reply is reviewed and signed by the clinician. Track errors found in review.
- Tell patientsExplain what is recorded or drafted by AI, and how to opt out.
- Check the rules firstPrivacy, device rules and your own clinical governance, before any patient data goes in.
More data: AI in Healthcare
More charts, each checked against its source, are on AI in Healthcare statistics. Dated events are in The AI Record.
How this page is made
Not yet in search engines. Health, legal and finance pages wait for a named domain reviewer.
The adoption figure is the US Census Bureau’s own estimate, read from its download; the use cases are rated from published studies by fixed rules (pilot, rules v0.3). Editorial owner: AI Adoption editors. Disclosure: AI Adoption uses AI tools, including Claude from Anthropic, in its research and editorial work. Spot a mistake? Corrections.