Home · Personal · Health & Wellbeing · Mental wellbeing

PARTLY

As of 13 August 2026, AI can only partly challenge negative thoughts.

This still needs a person who signs their name to it.

Can you do it?

5 minutesto a draft.

15 minutesto something you’d act on.

Cost, all in£0

Skill needednone

Who has to check ityou

What the alternative costsA purpose-built alternative is Wysa, which offers AI wellbeing support with evidence-based self-help exercises.

If this goes wrong: the chatbot dismisses a serious concern or encourages you to argue with yourself when you need human support, and your distress may increase.

What to actually do

  1. Hand it to a person

    The route this page recommends

    A person who owns the outcome does this end to end, worth it when the failure is dear.

  2. Use a tool built for this

    Second choice

  3. Do it yourself

    The distant third

    A chat interface, no skill needed, and roughly 15 minutes until you can act on the result.

    How to actually do it

    1. Open ChatGPT, Claude or Gemini and paste the prompt.
    2. Describe one recent situation and the exact negative thought in one or two sentences when the chatbot asks.
    3. Add the emotion, its intensity, what you did next, and the observable facts that support or contradict the thought.
    4. Ask the chatbot to show which parts are facts and which parts are interpretations before accepting a balanced alternative.
    5. Rewrite the suggested alternative in your own words so it is believable and does not dismiss a genuine problem.
    6. Choose one small safe next step from the exercise and write down whether it helped after you have tried it.
    7. If the conversation mentions immediate danger, self-harm, harm to someone else, or urgent or worrying symptoms, stop using the chatbot and contact NHS 111 or use 111.nhs.uk; call 999 in an emergency.

    Prompt

    Help me challenge one negative thought using a cautious, non-clinical CBT-style exercise. I am in the UK. Do not make a clinical assessment, claim that the thought is irrational, or tell me to ignore serious concerns. Ask me for one situation and thought at a time, then help me separate observable facts from interpretations, identify evidence for and against the thought, and write a balanced alternative that is believable rather than artificially positive. Ask what emotion and behaviour the thought is affecting, and suggest one small safe next step. If I mention immediate danger, thoughts of harming myself or someone else, or symptoms that feel urgent or worrying, stop the exercise and tell me to contact NHS 111 or use 111.nhs.uk, and to call 999 in an emergency. Start by asking me to describe the situation in one or two sentences.

    Open it prefilled in ChatGPT or Claude, or copy it into Gemini, which takes no prefill link.

What it gets wrong

  • Cannot tell whether a negative thought reflects a real danger, abuse or another serious situation.
  • Cannot provide a clinical assessment of depression, anxiety or any other mental health condition.
  • Cannot reliably notice changes in your speech, behaviour or safety that a clinician or trusted person might recognise.
  • Cannot take responsibility for the consequences if its reframing is wrong or increases your distress.
  • Cannot replace treatment, safeguarding or urgent human support.

What caps this at PARTLY: stakes of error, judgement under ambiguity and regulated advice.

How we scored this

Five axes, each scored nought to two by hand: ten means AI carries the task cleanly, and the thresholds that turn a total into YES, PARTLY or NO are published in the methodology. Each axis name links to its definition.

AxisScore (0–2)
Output2
Inputs2
Verification1
Liability0
Effort delta2
Total7 / 10

The methodology and its thresholds are published in full.

FAQ

Can ChatGPT help me challenge negative thoughts?
Yes, it can guide you through a structured exercise that separates facts from interpretations and produces a more balanced alternative thought. Treat the result as a reflection aid, not a clinical assessment or substitute for professional support.
Is it safe to use AI for negative thoughts?
It can be useful for ordinary reflection, but it is not reliably safe for serious, urgent or ambiguous situations. If you feel in immediate danger or have thoughts of harming yourself or someone else, contact NHS 111 or use 111.nhs.uk, and call 999 in an emergency.
Can AI tell me if my negative thoughts are irrational?
No. A chatbot cannot reliably know whether a thought is inaccurate, a response to a real risk, or connected to a mental health condition. Ask it to distinguish facts from interpretations instead, then discuss serious or persistent concerns with a qualified professional.
Should I use Wysa to challenge negative thoughts?
Wysa is a purpose-built wellbeing app offering AI support and evidence-based self-help exercises, so it fits this task more directly than a general chatbot. It still cannot provide a clinical assessment, assess urgent risk reliably or take responsibility for your care.

Nearby answers

All Mental wellbeing rulings

Assessed by gpt-5.6-luna (gpt-5.6-luna) on 2026-08-13, second-checked by an independent model. Wrong somewhere? Email [email protected] and it gets re-checked.

The newsletter

AI news, new answers and product picks, straight to your inbox.