Human-centred AI · Evidence-aware guide

AI Prompts for Reflection, Not Diagnosis

Technology can make an invisible pattern easier to see. It cannot decide what that pattern means on its own. AI can help structure reflection and questions, but it should not assign a diagnosis, interpret risk alone or present generated certainty as professional judgement.

Reviewed 24 August 2026British English · UK contextExplore Human-Centred AI and Digital Wellbeing
Direct answer

Use the information. Keep the conclusion proportionate

AI can help structure reflection and questions, but it should not assign a diagnosis, interpret risk alone or present generated certainty as professional judgement.

Outputs depend on the prompt, model, training data and missing context. Fluent language can make an uncertain or incorrect interpretation appear authoritative.

A useful next step

Use AI to organise observations, generate questions or compare options, then verify consequential information against reliable sources and appropriate professionals.

Four questions

Question. Context. Limits. Action

A compelling label, score or explanation may feel complete. A sound decision also needs context, limits and a clear account of what the information can change.

01

What question are you really trying to answer

AI can help structure reflection and questions, but it should not assign a diagnosis, interpret risk alone or present generated certainty as professional judgement.

02

What changes the meaning

Outputs depend on the prompt, model, training data and missing context. Fluent language can make an uncertain or incorrect interpretation appear authoritative.

03

What the information cannot establish

Do not enter identifiable health information without understanding data handling. Immediate safety concerns require human and emergency routes, not an AI conversation.

04

What action is proportionate now

Use AI to organise observations, generate questions or compare options, then verify consequential information against reliable sources and appropriate professionals.

Practical sequence

Describe. Verify. Decide. Review

The sequence supports informed choice while keeping wellbeing information in its proper place. It does not turn an observation into diagnosis or treatment.

01

Describe what is happening before explaining it

Record the experience you can observe, when it happens, how often it appears and what it changes in everyday life. Keep what happened separate from the explanation that currently seems most persuasive.

02

Verify the information that matters

Use a current NHS, NICE, WHO, GOV.UK or other appropriate primary source. Then ask whether the information concerns the same population, intervention and purpose. Similar words can conceal important differences.

03

Choose a route that matches the concern

Use AI to organise observations, generate questions or compare options, then verify consequential information against reliable sources and appropriate professionals.

04

Decide how the plan will be reviewed

Define what improvement, persistence, deterioration, adverse effect or new information would change the plan or require a referral. Otherwise, review can become an impression rather than a decision.

Scope and safety

Do not enter identifiable health information without understanding data handling. Immediate safety concerns require human and emergency routes, not an AI conversation. iMetaWellness provides complementary wellbeing information and support within stated roles. It does not diagnose, prescribe treatment or replace a GP, NHS service or appropriately regulated professional. Use NHS 111 for urgent advice and call 999 where there is immediate danger.

In one sentence

What this guide changes in practice

AI can help structure reflection and questions, but it should not assign a diagnosis, interpret risk alone or present generated certainty as professional judgement.

Use AI to organise observations, generate questions or compare options, then verify consequential information against reliable sources and appropriate professionals.

The boundary that matters most

Do not enter identifiable health information without understanding data handling. Immediate safety concerns require human and emergency routes, not an AI conversation.

Uncertainty is useful when it leads to a clearer question or a more appropriate professional route. It is not a reason to make the claim sound stronger.

Private support

Begin with the evidence and then consider the person

Use the patient form if you want to ask about suitability or scope. When the appropriate appointment is already clear, you can book an available time directly.