FCA authorised · FRN 724952 0117 325 0027Quote & buy →
Aipedia: Doctors and medical practices
The prescriber remains responsible for every prescription they sign, whether an AI safety check flagged a problem, missed it or was overridden. Medication errors can lead to claims, so how you respond to alerts matters.
Part of: AI in doctors and medical practices
In short
Prescribing systems increasingly use AI to flag interactions, dose errors, contraindications and overdue monitoring, and to find patients at risk across a practice list. The GMC applies Good medical practice to these tools: competence (paras 3 and 13), accurate records (para 69) and reporting adverse incidents with software (para 73). The GMC's separate prescribing guidance also applies. The main risks are alert fatigue, wrongly overridden warnings and over-reliance on a silent system. Responsibility may be shared between prescriber, practice and supplier, but one defence organisation's published position on AI excludes product liability and system defects.
Traditional prescribing alerts follow fixed rules. Newer tools add AI to rank alerts by likely significance, predict which patients are at risk of harm from a medicine, spot dosing patterns that look wrong for a patient's weight or kidney function, and search a practice list for patients who have missed monitoring blood tests. In hospitals, AI may support pharmacist review of drug charts. Practices also use tools that suggest medication review priorities or draft repeat prescription decisions for a clinician to approve.
The government's response to the National Commission frames reporting product problems as part of a professional's "duty of care" (Rec 34).
Illustrative scenario: a GP adds an antibiotic for a patient on a long-term medicine with a known serious interaction. The system shows several alerts, most of which the GP sees many times a day and routinely dismisses. The important warning is overridden with the rest. The patient is later admitted with a predictable adverse reaction, and the record contains no reason for the override.
Alert fatigue is one well-known failure: when tools produce too many low-value warnings, clinicians stop reading them. The reverse risk is a tool that misses a problem because of incomplete data, such as medicines prescribed elsewhere or an out-of-date kidney function result, and a prescriber who assumes silence means safety. Monitoring tools can also fail quietly if a search is set up wrongly, leaving patients unmonitored for months.
AI that drafts repeat prescription decisions creates a further risk. If a clinician approves a batch of AI-recommended repeats without opening each record, a medicine that should have been reviewed or stopped can continue. The GMC research found doctors worried about over-reliance and deskilling, and prescribing is where those risks turn quickly into harm. Patients should also be told when AI has informed a decision about their medicines, as the GMC expects.
The prescriber is responsible for the prescription and is the usual defendant. The practice or trust is responsible for choosing, configuring and monitoring the system, and NHS England's ambient scribing guidance makes deploying organisations responsible for clinical safety cases, a model the government has accepted for other AI (Rec 31). The supplier may be liable under contract or, if AI counts as a product, under the Consumer Protection Act 1987, which one defence organisation says is unclear. For NHS work, CNST or CNSGP usually handles the claim. For private work, the prescriber's own cover responds, subject to its terms, and one defence organisation's published position on AI excludes product liability and system failures or defects. Rec 28 of the government response is examining how contracts allocate responsibility. See AI liability.
See also AI for lab results, chatbots for clinical questions, does medical indemnity cover AI and the doctors AI hub.
Apex arranges indemnity for doctors and clinics, including private work outside NHS schemes. A broker reviews your cover for private prescribing and how your policy treats AI-assisted medication decisions.
You may be, if a reasonably competent prescriber would not have overridden it. Recording a clinical reason for overriding a significant alert is one of the most useful things you can do to defend the decision later.
Possibly, under contract or product liability, but this is unsettled. One defence organisation argues AI is not clearly a product under the Consumer Protection Act 1987, and its published position excludes product liability and system defects, so the prescriber is usually pursued first.
Yes. GMP para 73 requires doctors to report adverse incidents involving medical devices, including software and digital tools, and para 75 to raise concerns about systems. The government response treats reporting product problems as part of a professional's duty of care.
Tell us about your practice and the work you do. A broker will review how your current cover treats AI-assisted work, or arrange new cover. No fee for a quote. Or call 0117 325 0027.
Start an indemnity 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.
