One in seven people skip their GP and turn to AI chatbots for health advice
Fifteen percent of the public have used AI chatbots for health advice instead of contacting a doctor or NHS service, according to research by King's Health Partners, Responsible AI UK, and the Policy Institute at King's College London. Ten percent have used AI for mental health support instead of seeing a trained professional.
The shift raises safety concerns. One in five people who sought health advice from AI say the technology discouraged them from seeing a professional. Another 21% decided against seeking professional care because of what an AI chatbot told them. Recent evidence shows AI chatbots misdiagnose in up to 80% of early medical cases.
Why people are choosing AI over their doctor
Convenience tops the list. Forty-six percent cite ease of access, while 45% say curiosity drives their use. Thirty-nine percent use AI because they're unsure if their concern warrants a GP visit. A quarter point to long NHS waiting times.
Among those who have used AI for health advice, 59% report it has been good for their physical health and 53% for their mental health. But the public holds a different view of AI's broader impact-42% believe AI chatbots are bad for mental health overall, compared with 31% who think they're good.
Public demand for regulation far outpaces current safeguards
Three-quarters of the public want AI tools used in patient care to be officially approved and regulated, even if approval slows adoption. Just 17% support doctors choosing AI tools freely without oversight.
The reality is starkly different. The UK has no single regulatory framework for AI for healthcare. The National Commission into the Regulation of AI in Health Care is now examining how to close this gap, which critics including the Nuffield Trust and Royal College of Physicians have called a "wild west" of adoption.
Majorities of the public-between 58% and 63%-say they should be told in advance and given the option to opt out of AI being used in their care for reading test results, reviewing X-rays, listening to appointments, and deciding queue priority. This preference holds across age groups and gender.
Anxiety about safety dominates public sentiment
Anxiety about safety and accuracy is the top emotion 39% of the public feel about NHS use of AI for clinical tasks. Overall, people are twice as likely to feel negative emotions (63%) as positive ones (28%) about the prospect.
A stark gender divide runs through the data. Forty-six percent of women feel anxious about NHS AI use compared with 31% of men. Women are also far more likely to oppose AI in clinical decision-making (46% versus 30%) and uncomfortable with GPs using AI chatbots in consultations (65% versus 45%).
Young people also show higher skepticism. Half of 18-to-24-year-olds oppose clinical use of AI in the NHS, compared with 36% of those 65 and over. One in four people in this age group report AI has been bad for their mental health-the highest of any age group.
Who gets blamed when AI makes a mistake
The public holds doctors accountable first. If AI misses a health problem in a clinical image, 34% would primarily blame the treating doctor or health care professional. Twenty-four percent would blame the NHS trust that deployed the tool. Just 6% would hold the AI developer primarily responsible.
When asked who should bear responsibility, the public's expectations shift. A majority (55%) say a second doctor should review before any decision if an NHS-approved AI system disagrees with a doctor's diagnosis. Only 7% are willing to follow the AI's advice based on it potentially being more accurate.
The public overestimates how much AI doctors already use
People guess that 39% of GPs use AI in clinical decision-making. The actual figure is 8%.
Trust in doctors remains higher than trust in AI across common clinical scenarios. For psychological therapy-where trust in doctors is strongest-46% say they trust a doctor much more than AI, compared with just 1% who trust AI more.
That trust erodes when context matters. When doctors are described as being at the end of a long and busy shift, trust in them falls sharply. For psychological therapy, the proportion saying they trust a doctor much more drops from 46% to 25%. For skin cancer detection, it falls from 30% to 16%.
Data privacy concerns persist
Forty-seven percent of the public feel uncomfortable with their NHS health records being used to train AI if the data could identify them personally. Comfort rises to 40% when data would not identify them, though 36% remain uncomfortable even in that scenario.
Your membership also unlocks:
