FCA authorised · FRN 724952 0117 325 0027Quote & buy →
Aipedia: Translators and interpreters
Machine translation of patient information can introduce dosage, consent or allergy errors that harm patients, so medical text needs full human translation or revision by a qualified linguist. NHS England has raised patient-safety concerns about AI translation apps.
Part of: AI in Translators and interpreters
In short
Translators and agencies translate patient leaflets, consent forms, discharge letters and clinical correspondence, sometimes with MT or a generative AI chatbot in the workflow. NHS England's improvement framework of 27 May 2025 raised concerns about AI translation apps widely used across the NHS, citing accuracy and patient-safety risks. An error in a dose, a warning or a consent explanation can lead to harm. The clinician or NHS body usually carries clinical liability to the patient, but may seek to recover from the translator whose error caused it. Translators' PI usually responds to such claims, subject to the wording and any exclusions.
Healthcare providers, pharmaceutical companies and research teams commission translation of patient information leaflets, consent forms, clinical letters, appointment information and trial documents. Some arrive pre-translated by MT for post-editing. Some translators use AI tools for terminology or consistency checks. Separately, clinicians and patients use translation apps directly at the point of care, which is the use NHS England focused on.
Medical text also has to work for its reader. A leaflet written for patients uses plain language and a particular reading level, and a consent form must explain risks in a way a patient can weigh up. Engines tend to reproduce the register of the source, or slip into technical terms, so a translation can be accurate word for word and still fail the patient it was written for.
NHS England published its Improvement framework: community language translation and interpreting services on 27 May 2025. It raises concerns about "AI translation apps that are currently widely used across the NHS", accepts they are convenient but carry accuracy and patient-safety risks, and recommends a national policy briefing on ethical and appropriate AI use, covering clinical safety, when AI is suitable and when alternatives should be prioritised. That briefing had not been confirmed as published when this guide was written.
A translator owes the client reasonable care and skill under section 13 of the Supply of Goods and Services Act 1982. CIOL's AI Voices white paper says high-stakes translation needs humans "in the loop", and that AI errors are more frequent in less common languages, which are often the languages patients need. Patient information is health data, so a translator working for an NHS body or clinic usually acts as a processor under Article 28 of UK GDPR, and the ICO's AI guidance applies to any tool used.
Illustrative scenario: a clinic sends a translator a machine-translated aftercare leaflet to "check quickly". The engine rendered a twice-daily dose as twice weekly in one paragraph and dropped "do not" from a warning about alcohol. The translator, working to a short deadline, misses both. A patient follows the leaflet and is harmed. The patient claims against the clinic, which seeks a contribution from the translator.
Errors that matter most in medical text include:
The patient will usually claim against the clinician, practice or NHS body, which carries clinical liability. That organisation may then seek to recover some or all of its loss from the translator whose error contributed. Translators' PI usually responds to such a claim if it alleges negligent professional work, subject to the wording, the limit and any exclusions; some wordings treat bodily injury claims differently, so check. Dental and medical teams using translation tools have their own considerations, covered in AI translation in dental consultations. For contract and evidence work, see AI in legal and contract translation.
These controls help keep medical translation safe and defensible:
Apex arranges PI for translators who work on medical and patient information. A broker checks your limit, any bodily injury wording and how the policy treats AI-assisted work. See our professional indemnity for translators and interpreters page, or return to the AI in translation and interpreting hub.
It has not issued a ban. Its improvement framework of 27 May 2025 raised accuracy and patient-safety concerns about AI translation apps widely used in the NHS and recommended a national policy briefing on when AI is appropriate. NHS buyers may reflect this in contracts.
Usually the clinician or healthcare provider, who owes the patient a duty of care. The provider may then seek a contribution from a translator whose negligent error contributed to the harm. The outcome depends on the facts and the contracts between the parties.
Translators' PI usually responds to claims alleging negligent professional work, including where the loss follows injury, but some wordings handle bodily injury differently or exclude it. Ask your broker how your policy treats injury that follows a translation error.
Tell us about your firm and the work you do. A broker will review how your current PI treats AI-assisted work, or arrange new cover. No fee for a quote. Or call 0117 325 0027.
Start a PI quote Request a callbackApex Insurance Brokers Limited is authorised and regulated by the Financial Conduct Authority. Registered in England and Wales, company number 07014570. This page is general information about how AI is used in the profession and how insurance may respond. It is not legal, regulatory or insurance advice on your circumstances, and it does not guarantee that cover will be available or on what terms.
