There’s no shortage of hype about artificial intelligence in medicine. In general practice specifically, I think the realistic and valuable role over the next few years is narrower — and more useful — than the headlines suggest: taking administrative weight off clinicians so they can spend more time with patients.
Augmentation, not automation
The clearest example is documentation. A large part of a GP’s day goes into notes, referrals, and paperwork. Tools that help draft documentation from a consultation — with the doctor reviewing and approving every word — can give time back without touching clinical judgement. The key phrase is “with the doctor reviewing”: the AI drafts, the clinician decides.
That principle is the line I think general practice has to hold. AI can summarise, surface, and speed things up. It should not be making diagnoses or treatment decisions on its own, and patients are entitled to know that a human doctor is responsible for their care.
Privacy is non-negotiable
Health information is among the most sensitive data there is. Any AI tool used in a clinic has to meet a high bar for how patient data is stored, secured, and used — and that bar should be set before the tool is deployed, not bolted on afterwards.
Used with those guardrails, the upside is real: less administrative grind, less burnout, and more of the time that drew people to general practice in the first place. Used carelessly, it’s a liability. The job of anyone running clinics is to capture the first while refusing the second — practical, supervised, and privacy-first, rather than chasing whatever’s loudest this month.