Geo News
Community curated by people like you
LatestAICryptoHealthWorld AffairsUS Politics
AI medical scribes making dangerous errors in NHS patient records, watchdog warns
00

AI medical scribes making dangerous errors in NHS patient records, watchdog warns

Aug 31, 2026

Healthwatch England has warned that AI medical scribes used across NHS England are making critical errors, such as misrecording diagnoses and drug names. Despite 27 different AI scribes being deployed to reduce administrative burdens, the Medicines and Healthcare products Regulatory Agency has not classified them as medical devices, leaving them without national safety testing. Patient advocates and clinicians warn that these errors threaten patient safety, placing the burden of detection on patients themselves.

AI scribe transcription errors

  • ▪An AI scribe falsely recorded that London GP Dr. Shier Ziser Dawood told a patient to continue taking Prozac, despite the drug never being prescribed or discussed.
  • ▪An AI-generated summary letter in NHS England omitted a hospital consultant's instruction that a patient should seek a repeat prescription for migraine medication.
  • ▪An AI scribe used in an NHS England hospital incorrectly recorded a patient's scan result as 'demyelination' instead of 'null demyelination,' dropping the word that reversed the diagnosis.
  • ▪An AI scribe used in NHS England confused a GP-prescribed drug with a different drug of a similar name, which was subsequently identified by the patient.

Patient safety risks

  • ▪Healthwatch England reported hearing multiple stories from patients who noticed AI scribe errors that healthcare professionals had missed.
  • ▪Healthwatch England warned that transcription errors by AI scribes could persist in permanent medical records and affect patient care if patients do not catch them.
  • ▪Research by Dr. Charlotte Blease of Uppsala University indicates GPs believe AI scribe errors are more likely during consultations involving multiple people, complex medical histories, or non-English speakers.

Medical device regulatory gaps

  • ▪GPs and hospital doctors in England are currently using 27 different AI scribes without centralized regulatory oversight.
  • ▪Medicines and Healthcare products Regulatory Agency guidance published in August 2026 states that passive transcription systems are not classified as medical devices, whereas systems suggesting diagnoses or treatments may be.
  • ▪The Medicines and Healthcare products Regulatory Agency decided not to classify AI scribes as medical devices, leaving them without England-wide safety and effectiveness testing.

Clinical workflow time pressures

  • ▪A survey of 1,003 UK GPs by Dr. Charlotte Blease in 2025 found that more than half believed ambient AI records were more accurate than their own notes.
  • ▪London GP Dr. Shier Ziser Dawood warned that AI scribes do not yet save time because doctors must review all transcripts to check for errors.
  • ▪The UK government's 10-year health plan for NHS England expects AI scribes to reduce administrative burdens and free up staff time for patient care.

Patient advocacy concerns

  • ▪Healthwatch England recommended that patients should be informed when an AI scribe is used and provided with their notes to check for errors.
  • ▪Rachel Power, chief executive of the Patients Association, stated that trust in AI scribe technology is missing due to a lack of good communication and genuine partnership with patients.

2 sources

The Guardian
Doctors’ AI scribes get names of drugs and diagnoses wrong, NHS watchdog warns
View source article
Thenextweb
AI scribes are getting drug names and diagnoses wrong in NHS records
View source article

Featured stories

View more in AI assistants & chatbots

Apple accuses OpenAI of destroying evidence in trade secrets lawsuit over former engineer

Aug 31, 2026 · 7 sources

EU designates ChatGPT, Reddit and Roblox under strictest Digital Services Act rules

Aug 31, 2026 · 4 sources

Bank of England governor warns AI models threaten global financial stability

Aug 31, 2026 · 4 sources

Anthropic warns Claude users after infostealer malware hijacks active login sessions

Aug 30, 2026 · 4 sources

Story comments

Loading comments…

Related entities

United Kingdom

Topics

AI assistants & chatbotsAI safety & social impactAI regulation & lawsuitsPatient safety

Featured stories

View more in AI assistants & chatbots

Apple accuses OpenAI of destroying evidence in trade secrets lawsuit over former engineer

Aug 31, 2026 · 7 sources

EU designates ChatGPT, Reddit and Roblox under strictest Digital Services Act rules

Aug 31, 2026 · 4 sources

Bank of England governor warns AI models threaten global financial stability

Aug 31, 2026 · 4 sources

Anthropic warns Claude users after infostealer malware hijacks active login sessions

Aug 30, 2026 · 4 sources