Why am I asking AI about my health instead of speaking to a doctor?
It’s 3am, and I’m asking a machine about my health
There’s a kind of worry that arrives at 3am. One thought becomes five, five become 17, and suddenly you’re diagnosing yourself with something you hadn’t even heard of yesterday. Sleep is out of the question, so you reach for your phone, hoping to find an explanation that will let you get back to sleep.
I’m 51, lift weights and go to the gym at least three times a week. I try to look after myself, although my perimenopausal body sometimes has other ideas. When something feels wrong, I want to understand why immediately. I don’t necessarily need a diagnosis, but I do need somewhere to start without feeling embarrassed or foolish. Artificial intelligence offers exactly that. It answers immediately, never sounds irritated and doesn’t mind another question. Unfortunately, being available doesn’t make it medically qualified.
AI offers answers without appointments, telephone queues or awkward conversations. But there’s a big difference between getting an answer and getting the right medical advice.
When getting a GP appointment is tough
Getting medical advice isn’t always easy. My GP surgery uses an online system where I describe my symptoms and answer questions before anyone decides what happens next. It reminds me of those children’s adventure books where you choose the ending. Turn to page 35 for one outcome, page 68 for another. Except choosing the wrong route when you’re worried about your health isn’t as entertaining.
Sometimes the system works, but at other times the questions don’t fit what I’m trying to explain. Then comes the wait for a callback, which usually arrives during a Teams meeting. Meanwhile, I’m wondering whether to try again tomorrow or hope the problem just disappears overnight. Compare that with AI. No appointments, no telephone queues and no explaining your symptoms to three different people. You get an answer that sounds confident, whether it’s correct or not.
A Confused.com survey found that 42% of respondents identified speed compared with waiting for a doctor’s appointment as an advantage of AI. That doesn’t establish that NHS delays caused people to use it. It does help explain the appeal: when you’re worried, waiting for an answer can feel much longer than it is.
The survey describes why AI appeals. It doesn’t prove that its answers are clinically reliable.
Are people really using AI to diagnose themselves?
A Confused.com press release about its OnePoll survey reports that it questioned 2,000 people in November 2025. Its release says 59% used AI for self-diagnosis and lists symptom checks, side-effect questions and wellbeing advice among the uses. Those activities aren’t interchangeable: researching a symptom doesn’t necessarily mean deciding what illness you have.
This is commissioned consumer research, not evidence that AI diagnoses illness accurately. The public release doesn’t provide the full questionnaire or detailed respondent bases for every figure, so I wouldn’t turn its headline into an unqualified statement about all Britons.
For women in midlife, making sense of menopause symptoms can be difficult. Fatigue might involve hormones, poor sleep, iron deficiency or stress. Throw work, children, ageing parents and household responsibilities into the equation, and even finding time to investigate becomes another job. I’ve often wondered whether something is perimenopause or the consequences of trying to do everything at the same time. Having somewhere to ask questions can feel helpful when you’re struggling to work out which symptoms are important.
Can you trust AI for medical advice?
I don’t think using AI for health information automatically makes someone irresponsible. There are perfectly reasonable reasons to ask questions. You might want to understand medical terminology, prepare for an appointment or work out how to explain symptoms you’ve struggled to describe. For me, its greatest advantage is helping organise my thoughts before speaking to someone qualified. Sometimes I want to understand enough about a problem to have a conversation with my GP.
But there are limitations. A chatbot can’t examine me, assess my physical condition or establish what’s happening through appropriate clinical tests. It can misunderstand symptoms, overlook important information or offer reassurance when medical attention is necessary. There’s also the risk of feeding anxiety. One question leads to another until you’ve constructed an entire medical history around something that might have an ordinary explanation.
There’s evidence behind that concern. An Oxford-led study published in February 2026 tested nearly 1,300 people using chatbots in medical scenarios. They didn’t make better decisions than people using conventional approaches, such as web searches or their own judgement. This was a scenario-based study, not a test of every chatbot or every consultation. Still, sounding knowledgeable and helping someone make a safe decision are different achievements.
And what happens to the information we’re sharing? Medical histories, symptoms and medication details are sensitive. Before entering them into any AI service, we need to understand how that information might be stored or used. Convenience shouldn’t mean abandoning common sense or privacy when we’re discussing details we might hesitate to share with another person.
What happens in the gap before an appointment?
We hear plenty about whether artificial intelligence will eventually replace healthcare professionals. I’m more interested in what happens before someone gets to see one. The NHS provides care that many of us value enormously, but accessing that care can still involve frustrating delays. I don’t expect a GP to answer my questions at 3am. I do, however, want a clear route to professional advice when I need it.
AI can help someone prepare questions or understand information, but it shouldn’t determine whether they seek urgent care. A reassuring response should never override symptoms that need medical assessment. I want help putting my questions in order, then a qualified person to assess the answers.
The irony of our relationship with screens
The strangest part is how we treat our children’s relationship with technology. We complain about excessive screen time, gaming and children reaching for devices instead of engaging with people. Yet when we’re frightened or uncertain, many of us do the same. The difference is that our questions sound more grown-up, and we can justify our screen time as something useful instead of entertainment.
At 3am, I might ask a machine to help make sense of something worrying me. By morning, though, I’d still prefer to speak to a healthcare professional who can listen, examine me and take responsibility for clinical decisions. Technology has its uses, but there’s something slightly unsettling about telling our children to put their screens away while asking ours whether we’re going to be all right.
If you need help: This is personal commentary, not individual medical advice. In England, NHS 111 can direct you to appropriate care when you need medical help now and can’t reach your GP. For a life-threatening emergency, call 999. Don’t wait for a chatbot to reassure you.
Source note: Confused.com’s commercial press release reports a OnePoll survey of 2,000 people in November 2025. It is not a clinical study. The full questionnaire and detailed subgroup bases were not available in the release reviewed. This article also draws on the Oxford-led medical-scenario study linked above.
More from Saffron & Cyrus
Stories, experiments and questions from Saffron & Cyrus.
Discover more from Saffron & Cyrus
Subscribe to get the latest posts sent to your email.
Saffron & Cyrus is an independent women’s editorial publication with a recognisable voice, real-life perspective, a defined midlife audience and a commercial model built around useful, relevant partnerships.

Author
https://www.facebook.com/aranda.rahbarkouhi/posts/pfbid031sjutgmHf2CKEBcRzNFBqVM3iRKk2Fn8Z2SQQE9Bf5zJQs5E7Pv9T5gttSWiEUQGl