FCA authorised · FRN 724952 0117 325 0027Quote & buy →
Aipedia: Pharmacists and pharmacies
You can use a general AI chatbot to help think through a medicines question, but you must not rely on its answer without checking it against an authoritative source. If an unchecked answer harms a patient, it will be hard to defend.
Part of: AI in pharmacists and pharmacies
In short
Pharmacists may ask general generative AI chatbots about doses, interactions, compatibility, off-label use and renal or paediatric adjustments. These tools can produce confident answers with invented doses or references. The GPhC position statement says pharmacists must understand the limitations of a tool, check outputs for accuracy and misleading content, and remain personally accountable. Standards 4 and 5 require knowledge and professional judgement, and Standard 7 restricts entering patient data. A chatbot answer is not a recognised reference source, so advice based on it is usually judged as if the pharmacist gave it without checking.
Typical uses include a quick check on an interaction, a starting point for an unusual dose calculation, a summary of evidence on an off-label use, a draft answer to a prescriber's query, or plain-English wording for a patient. Some pharmacists use a chatbot as a sounding board when the usual references do not answer the question directly. General chatbots are not designed or validated as medicines information resources, and the answer can change if the question is phrased differently.
These duties apply whether or not AI is involved:
Guidance from a pharmacists' defence association, reported by the Pharmaceutical Journal, says pharmacists should critically evaluate outputs before acting and should not use AI where it adds uncertainty or reduces decision clarity.
Illustrative scenario: a hospital ward asks a pharmacist about dosing an unfamiliar medicine for a patient with reduced kidney function. The pharmacist asks a generative AI chatbot, which gives a dose adjustment and cites a guideline. The pharmacist passes the dose on. The guideline exists, but it does not say what the chatbot claimed, and the dose is too high for the patient's kidney function. The patient suffers toxicity.
Hallucinated references, confident but wrong doses, mixed-up units, out-of-date information and answers based on another country's products are all known failure modes of general chatbots. The more specialised the question, the more likely the answer is to be thin or invented, which is the opposite of when pharmacists most want help. Doctors face the same issue; see doctors and chatbots for clinical questions.
In a claim, the question is whether a reasonably competent pharmacist would have given that advice. "A chatbot said so" does not answer it. A pharmacist who checked a recognised reference and reached a reasonable view is in a much stronger position than one who passed on an unchecked output, even if both turned out to be wrong. PI usually responds to claims arising from advice given in the course of professional practice, subject to the policy terms, and how a policy treats advice drawn from AI depends on the wording. A pharmacists' defence association has warned, as reported in the Pharmaceutical Journal, that pharmacists risk becoming "liability sinks" because AI cannot be held legally liable. A regulator would also ask why the pharmacist did not verify the answer.
Practical controls include:
See also AI clinical checks, AI in hospital pharmacy, AI hallucination liability and the pharmacy AI hub.
Apex arranges indemnity for pharmacists in every sector. A broker reviews how your policy treats advice prepared with the help of generative AI and whether any AI exclusions apply.
The GPhC does not ban it, but you must understand the tool's limits, check its output for accuracy and remain accountable for any decision. Follow your employer's policy, and do not enter identifiable patient data into a tool that has not been approved.
Open it and check. Chatbots can cite real sources for claims those sources do not make. The GPhC requires pharmacists to check outputs for accuracy and misleading content, and relying on a citation you have not read is hard to justify.
Not unless the tool has been approved for patient data by your employer or owner. Standard 7 and the GPhC position statement require you to protect confidentiality and data. A question can usually be framed without identifiers, such as age range, kidney function and current medicines.
Cover usually depends on the advice being given in the course of your professional practice and on the policy wording, not on where the information came from. Some policies may contain AI-related exclusions, so check the wording and declare how you use AI.
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.
