TL;DR

The Pharmacists’ Defence Association has issued guidance on AI in pharmacy practice. Its central line is that a system may shape a professional judgement but cannot carry responsibility for one — accountability stays with the pharmacist. The association is deliberately neither endorsing nor discouraging the technology.

Where attention is needed

The guidance identifies the situations that warrant particular care: systems issuing recommendations, systems dealing directly with patients, systems that make the reasoning behind a decision less visible, and systems that alter clinical processes. Automation inside patient medication record software is given as the working example.

That third category is the sharp one. A tool producing a defensible answer through a process the pharmacist cannot inspect creates a specific problem — the professional remains answerable for a judgement they cannot fully reconstruct. The association’s position is blunt on this: where AI introduces uncertainty, muddies how a decision was reached, or weakens accountability, it should not be used at all.

Alongside that, the document covers how these systems work, the risks attached to different types, questions of regulation, data quality and bias, liability, and practical risk-reduction steps.

Consistent with the regulator

This is not a solo intervention. The General Pharmaceutical Council published a position statement in April holding that existing standards apply unchanged when AI is involved, and that pharmacy professionals stay personally accountable for their decisions and actions.

Members are being asked for their experiences through a survey closing on 20 September, which will feed the association’s future policy work.

Looking forward

Three UK bodies landed on the same principle within a day — the advertising regulator on evidencing AI claims, the FCA on consumers misreading what protection they have, and now pharmacy on where responsibility sits. None created new rules. Each said the existing framework already applies.

That is the pattern British professionals should take from this week. The regulatory answer to AI in the UK is arriving as clarification rather than legislation, which is faster but harder to notice. Waiting for an AI Act before reviewing your own accountability arrangements means waiting past the point at which you are already expected to have done it.