AI in Healthcare Statistics 2026: FDA Devices, Doctors, Hospitals and Patients
The FDA's list of AI-enabled medical devices held 1,614 authorised devices with decisions up to 29 June 2026. A record 335 were authorised in 2025, up 42.6% on 2024. Radiology accounts for 76.2% of them. In 2026, 81% of US physicians used AI in their practice, according to the AMA. About a third of US adults now use AI for health information, according to KFF and Pew.
Key statistics
- The FDA list of AI-enabled medical devices held 1,614 authorised devices with decisions up to 29 June 2026, calculated by StatisticsGlobal from US Food and Drug Administration data.
- FDA authorisations of AI-enabled medical devices rose 42.6% in 2025, from 235 in 2024 to 335, calculated by StatisticsGlobal from US Food and Drug Administration data.
- The share of US physicians using AI in their practice reached 81% in 2026, according to the American Medical Association.
- US non-federal acute care hospitals using predictive AI integrated into their EHR reached 71% in 2024, according to ASTP/ONC.
- US adults using AI chatbots for at least one of eight health reasons made up 34% in 2026, according to Pew Research Center.
FDA-approved AI medical devices statistics
The FDA publishes a list of AI-enabled medical devices authorised for marketing in the US. We counted every row of its CSV download by year of final decision.
Authorisations have risen almost every year since 2016. The FDA authorised 335 AI-enabled devices in 2025, more than in 2016 to 2019 combined (190). Growth averaged 24.1% a year from 2020 to 2025.
Growth is uneven. 2024 added only 4.0% on 2023, then 2025 jumped 42.6%. The first half of 2026 ran just 2.8% ahead of the first half of 2025, so 2026 may not match the 2025 jump. The list only runs to decisions on 29 June 2026, so the full-year figure is not known yet.
Most of these devices are not "approved" in the strict sense. 96.2% reached the market through the 510(k) route, which clears a device as similar to one already on sale. Only 21 went through full premarket approval. The list is young too. 85.6% of all entries were authorised from 2020 onwards.
Source: US Food and Drug Administration device list. Counted by StatisticsGlobal.
-
The FDA list of AI-enabled medical devices held 1,614 authorised devices with decisions up to 29 June 2026, calculated by StatisticsGlobal from US Food and Drug Administration data.
How we calculated it: Count of all data rows in the FDA CSV download (one row per authorisation). Latest decision date in the file: 29 June 2026.
-
The FDA had authorised over 1,600 AI-enabled medical devices for marketing in the US as of September 2026, according to the US Food and Drug Administration.
-
The FDA authorised 335 AI-enabled medical devices in 2025, the most in any year on its list, calculated by StatisticsGlobal from US Food and Drug Administration data.
How we calculated it: Rows in the FDA CSV with a Date of Final Decision in 2025. Next highest year: 2024 with 235.
-
FDA authorisations of AI-enabled medical devices rose 42.6% in 2025, from 235 in 2024 to 335, calculated by StatisticsGlobal from US Food and Drug Administration data.
How we calculated it: 335 devices with a 2025 decision date divided by 235 with a 2024 decision date, minus 1.
-
FDA authorisations of AI-enabled medical devices grew 24.1% a year on average between 2020 and 2025, calculated by StatisticsGlobal from US Food and Drug Administration data.
How we calculated it: Compound annual growth from 114 devices in 2020 to 335 in 2025: (335 / 114) to the power 1/5, minus 1.
-
The FDA authorised 181 AI-enabled medical devices in the first half of 2026, 2.8% more than the 176 in the first half of 2025, calculated by StatisticsGlobal from US Food and Drug Administration data.
How we calculated it: Rows with a decision date from January to June 2026 (181) against January to June 2025 (176). 181 / 176 minus 1 = 2.8%.
| Year of final decision | Devices authorised | Change vs previous year |
|---|---|---|
| 2016 | 18 | - |
| 2017 | 27 | +50.0% |
| 2018 | 65 | +140.7% |
| 2019 | 80 | +23.1% |
| 2020 | 114 | +42.5% |
| 2021 | 129 | +13.2% |
| 2022 | 162 | +25.6% |
| 2023 | 226 | +39.5% |
| 2024 | 235 | +4.0% |
| 2025 | 335 | +42.6% |
| 2026 (January to June) | 181 | +2.8% vs January to June 2025 |
AI medical devices by specialty
Each device on the FDA list has a lead review panel. That panel is the closest thing the list has to a medical specialty.
Radiology dominates. 1,230 of the 1,614 devices sit with the radiology panel. Most of these are imaging software. One of the FDA's own examples is deep-learning image sharpening.
The mix is not shifting much. Radiology took 254 of the 335 devices authorised in 2025, or 75.8%, close to its share of the whole list. Cardiovascular is a distant second at 9.5%, and neurology third at 4.5%.
The panel is the FDA's review route, not a full map of clinical use. A device can be used across specialties while being reviewed by one panel.
Source: US Food and Drug Administration. Calculated by StatisticsGlobal.
| Lead panel | Devices | Share of list |
|---|---|---|
| Radiology | 1,230 | 76.2% |
| Cardiovascular | 154 | 9.5% |
| Neurology | 73 | 4.5% |
| Anesthesiology | 30 | 1.9% |
| Gastroenterology-Urology | 27 | 1.7% |
| Hematology | 22 | 1.4% |
| All other panels | 78 | 4.8% |
| Total | 1,614 | 100% |
Physician AI adoption statistics
The American Medical Association has surveyed US physicians on AI every year since 2023. The 2026 wave covered nearly 1,700 doctors.
Physician AI use more than doubled in three years, from 38% in 2023 to 81% in 2026. Doctors also use it for more tasks. The average rose from 1.1 use cases per physician to 2.3.
The main uses are paperwork and reading, not diagnosis. Research summaries lead at 39%. Notes, charts and billing come next. Assistive diagnosis is last of the listed uses at 17%.
Doctors are using AI while worrying about it. 88% have at least some concern about losing clinical skills, and 85% want a say in how their organisation adopts AI. The AMA does not publish field dates or a margin of error, so treat small differences between uses with care.
Source: American Medical Association.
-
The share of US physicians using AI in their practice reached 81% in 2026, according to the American Medical Association.
-
The share of US physicians using AI in their practice was 38% in 2023, according to the American Medical Association.
-
Physician AI use in the US rose 43 percentage points between 2023 and 2026, from 38% to 81%, calculated by StatisticsGlobal from American Medical Association data.
How we calculated it: 81% in 2026 minus 38% in 2023, both from the AMA press release of 12 March 2026.
-
US physicians used AI for an average of 2.3 tasks each in 2026, up from 1.1 in 2023, according to the American Medical Association.
-
Summarising medical research and standards of care was the top physician use of AI in the US in 2026, at 39%, according to the American Medical Association.
-
AI for billing codes, medical charts or visit notes was used by 28% of US physicians in 2026, according to the American Medical Association.
-
Concern about AI-related skill loss was reported by 88% of US physicians in 2026, according to the American Medical Association.
-
Being consulted or involved in AI adoption decisions mattered to 85% of US physicians in 2026, according to the American Medical Association.
| Use | Share of physicians |
|---|---|
| Summaries of medical research and standards of care | 39% |
| Discharge instructions, care plans or progress notes | 30% |
| Billing codes, medical charts or visit notes | 28% |
| Chart summaries | 28% |
| Draft replies to patient portal messages | 19% |
| Translation | 18% |
| Assistive diagnosis | 17% |
Hospital AI adoption statistics
ASTP/ONC, part of the US Department of Health and Human Services, tracks predictive AI built into hospital electronic health records (EHRs). Its latest data covers 2024.
Seven in ten US hospitals used predictive AI inside their EHR in 2024, up from 66% in 2023. The fastest growth was in admin work. Billing use jumped from 36% to 61% of hospitals using predictive AI.
The gap between hospital types is wide. System-affiliated hospitals used predictive AI at 2.3 times the rate of independent hospitals in 2024. ASTP/ONC also found 90% of hospitals on the market-leading EHR used predictive AI, against 50% of hospitals using other EHR vendors.
This measures predictive models, such as risk scores and scheduling tools. It is not a measure of generative AI or chatbots, so it cannot be compared directly with the physician survey above.
-
US non-federal acute care hospitals using predictive AI integrated into their EHR reached 71% in 2024, according to ASTP/ONC.
-
US non-federal acute care hospitals using predictive AI integrated into their EHR stood at 66% in 2023, according to ASTP/ONC.
-
System-affiliated US hospitals used predictive AI at 2.3 times the rate of independent hospitals in 2024, 86% against 37%, calculated by StatisticsGlobal from ASTP/ONC data.
How we calculated it: 86% of system-affiliated hospitals divided by 37% of independent hospitals, both 2024, from ASTP/ONC Data Brief No. 80.
-
Among US hospitals using predictive AI, the share applying it to billing rose from 36% in 2023 to 61% in 2024, according to ASTP/ONC.
| Measure | 2023 | 2024 |
|---|---|---|
| Hospitals using EHR-integrated predictive AI | 66% | 71% |
| System-affiliated hospitals using predictive AI | 81% | 86% |
| Independent hospitals using predictive AI | 31% | 37% |
| Predictive AI users applying it to billing | 36% | 61% |
| Predictive AI users applying it to scheduling | 51% | 67% |
Patient use of AI chatbots for health
Two large surveys measured how many US adults use AI for health in 2026. KFF asked in February and March. Pew asked in June.
The two surveys agree. KFF found 32% of US adults turn to AI for health information or advice. Pew found 34% use chatbots for at least one of eight health reasons. The questions differ, so the 2-point gap is not a change over time.
Use falls sharply with age. Pew found 44% of adults aged 18 to 29 use chatbots for health, against 17% of those 65 and older. Speed is the top reason, followed by working out symptoms.
Privacy is the main worry. 77% of US adults are concerned about the privacy of medical information they give to AI tools, yet 13% of all adults have already entered personal medical information into one. Most users still check with a professional. 58% of physical health AI users later saw a doctor or provider, according to KFF. For which chatbots people use, see our AI chatbot market share statistics.
Source: Pew Research Center.
-
US adults turning to AI for health information or advice reached 32% in 2026, according to KFF.
-
US adults using AI chatbots for at least one of eight health reasons made up 34% in 2026, according to Pew Research Center.
-
US adults using AI for mental health information or advice in the past year stood at 16% in 2026, according to KFF.
-
US adults using AI chatbots to work out what is causing symptoms made up 25% in 2026, according to Pew Research Center.
-
US adults using AI chatbots to help understand lab results made up 20% in 2026, according to Pew Research Center.
-
US adults aged 18 to 29 were 2.6 times as likely as those 65 and older to use AI chatbots for health in 2026, calculated by StatisticsGlobal from Pew Research Center data.
How we calculated it: 44% of adults aged 18 to 29 divided by 17% of adults aged 65 and older who are chatbot health users, from the Pew chart in the same report.
-
US adults who used AI for physical health advice and later saw a doctor or provider made up 58% of those users in 2026, according to KFF.
-
US adults who had entered personal medical information into an AI tool made up 13% of all adults in 2026, according to KFF.
-
Concern about the privacy of personal medical information given to AI tools was reported by 77% of US adults in 2026, according to KFF.
| Reason | Share of US adults |
|---|---|
| To get health information quickly | 28% |
| To figure out what is causing symptoms | 25% |
| To get health information at little or no cost | 22% |
| To get information about treatments | 22% |
| To learn more about a diagnosis from a doctor | 22% |
| To help understand lab results | 20% |
| To get information they are uncomfortable talking about | 18% |
| To help decide whether to see a doctor | 15% |
| Any of these eight reasons | 34% |
How the sources measure AI in healthcare
The FDA list covers AI-enabled devices authorised for marketing in the US. The FDA says it updates the list periodically. Our counts use the CSV download, where the latest decision is dated 29 June 2026. One row is one authorisation, so a product updated several times can appear more than once. The FDA's own headline says over 1,600 devices as of September 2026, which matches our count of 1,614.
The AMA survey covered nearly 1,700 US physicians across specialties and practice settings. The AMA has not published field dates, sampling details or a margin of error for the 2026 wave.
ASTP/ONC uses the American Hospital Association Annual Survey IT Supplement. It covers non-federal acute care hospitals, with 2,080 responses in 2024. Use-case shares are among hospitals that use any predictive AI.
KFF polled 1,343 US adults from 24 February to 2 March 2026, with a margin of error of plus or minus 3 points. Pew surveyed 3,488 US adults on its American Trends Panel from 22 to 28 June 2026.
All five sources are US only. We found no comparable international data from a primary source for this page. These figures should not be read as global rates.
Frequently asked questions
How many AI medical devices has the FDA approved?
The FDA list of AI-enabled medical devices held 1,614 authorised devices with decisions up to 29 June 2026, counted by StatisticsGlobal. Most were cleared through the 510(k) route rather than full approval.
How many AI medical devices did the FDA authorise in 2025 compared with 2024?
The FDA authorised 335 AI-enabled devices in 2025, up 42.6% from 235 in 2024. In the first half of 2026 it authorised 181, 2.8% more than in the first half of 2025.
Which medical specialty has the most AI devices?
Radiology, by far. It accounts for 76.2% of AI-enabled devices on the FDA list, 1,230 of 1,614. Cardiovascular is second at 9.5%.
How many doctors use AI?
81% of US physicians used AI in their practice in 2026, up from 38% in 2023, according to the American Medical Association.
How many people use AI for health advice?
About a third of US adults. KFF found 32% turn to AI for health information or advice, and Pew found 34% use AI chatbots for at least one health reason.
Are people worried about privacy when using AI for health?
Yes. 77% of US adults are concerned about the privacy of personal medical information given to AI tools, according to KFF. Even so, 13% of all adults have entered personal medical information into one.
Sources
- US Food and Drug Administration, List of Artificial Intelligence-Enabled Medical Devices (CSV download), List page updated 4 September 2026, accessed 16 October 2026. Link · Saved data copy · Licence: US government work, public domain
- American Medical Association, AMA: AI usage among doctors doubles as confidence in technology grows, 12 March 2026. Link · Saved data copy
- ASTP/ONC (US Department of Health and Human Services), Hospital Trends in the Use, Evaluation, and Governance of Predictive AI, 2023-2024 (Data Brief No. 80), September 2025. Link · Saved data copy · Licence: US government work, public domain
- Pew Research Center, From diagnoses to treatments: why Americans use AI chatbots for health, 25 August 2026. Link · Saved data copy
- US Food and Drug Administration, Artificial Intelligence-Enabled Medical Devices, 22 September 2026. Link · Saved data copy · Licence: US government work, public domain
- American Medical Association, More than 80% of physicians use AI professionally: AMA survey, 12 March 2026. Link · Saved data copy
- KFF, KFF Tracking Poll on Health Information and Trust: Use of AI for Health Information and Advice, 25 March 2026. Link · Saved data copy
Changes to this page
- : Page published with FDA device data to June 2026, AMA and KFF and Pew 2026 surveys and ASTP/ONC 2024 hospital data.