One in seven UK adults using AI chatbots instead of seeing a GP

TL;DR:

• 15% of UK adults polled by King’s College London have substituted an AI chatbot for a GP visit, and a quarter of that group cite NHS waiting times as the reason.
• One in five who used a chatbot for health advice said the tool did not direct them to see a clinician; a similar share decided against a consultation based on what the chatbot told them.
• The study lands as a separate City Hall report has put a million London jobs in the path of AI disruption — healthcare is now visibly inside the consumer-AI substitution story, not adjacent to it.

The poll of more than 2,000 UK adults, analysed by researchers at King’s College London, is the first attempt to quantify British use of AI chatbots for health advice rather than measuring tool capability. Lead author Professor Graham Lord warned that public reliance on consumer chatbots is now creating “an unregulated AI healthcare system alongside the NHS”.

A regulatory gap, not a tool failure

The headline statistic — 15% chatbot substitution — sits beside two findings that point to safety, not adoption, as the bigger Resultsense concern. A fifth of chatbot users said the tool did not encourage them to seek a professional opinion, and a similar fifth then decided against a consultation because of what the chatbot had told them. That is the user-side risk pathway the Royal College of General Practitioners has been warning about for over a year.

Professor Victoria Tzortziou Brown, RCGP president, said it would be “highly concerning” if patients were turning to AI because of access barriers. Her statement is notable for separating two arguments often conflated in NHS-AI debates: AI tools can be useful in clinical settings under regulation, but they must not become a workaround for under-investment in primary care.

UK angle: the demand half of the AI-in-health debate

Most UK AI-in-healthcare coverage has focused on the supply side — NHS pilots, vendor contracts, the £330 million Palantir Federated Data Platform deal. The KCL study reframes the question. UK patients are already using AI for triage decisions, without clinical oversight, and the technology is filling a gap that NHS waiting times have created. Age splits in the data reinforce the policy challenge: 49% of 18-24s oppose clinical AI use within the NHS, against just 36% of over-65s, suggesting trust patterns that NHS England’s AI strategy has not fully accounted for.

Looking forward

Expect the Department of Health and Social Care to come under pressure to extend the General Pharmaceutical Council’s AI-in-pharmacy guidance approach into general practice, and for the MHRA to clarify which consumer chatbots fall inside its software-as-a-medical-device perimeter. The Regulating for Growth Bill set out in the King’s Speech is the most likely vehicle.