Home · Personal · Family & Relationships · Relationships & social life
As of 13 August 2026, AI can only partly recognise signs of relationship abuse.
Most people should hand this to a purpose-built tool.
Can you do it?
5 minutesto a draft.
30 minutesto something you’d act on.
Cost, all in£0
Skill needednone
Who has to check ityou
What the alternative costsThe supplied tool list gives no price for a specialist domestic-abuse support service.
If this goes wrong: the model minimises dangerous behaviour, labels an unclear situation too confidently, or suggests an action that increases your risk.
What to actually do
Use a tool built for this
The route this page recommends
Do it yourself
Second choiceA chat interface, no skill needed, and roughly 30 minutes until you can act on the result.
How to actually do it
- Open a private chat on a device and account your partner cannot access, if doing so is safe.
- Write a short timeline of specific incidents, including what happened, what was said, how often it happened and whether you felt frightened or controlled, while removing names, addresses and identifying details.
- Paste the timeline and the prompt into the chatbot, and ask it to separate observed facts from interpretations.
- Compare its explanation of possible patterns with current UK guidance from GOV.UK or the NHS, rather than treating the chatbot's label as a diagnosis.
- Ask the chatbot to turn the description into questions for a domestic-abuse adviser, GP or another trusted professional, without asking it to decide what you must do.
- If you are in immediate danger, stop using the chatbot and contact emergency services; otherwise contact a specialist domestic-abuse service or a trusted person using a safe method.
Prompt
Help me organise what I have experienced so I can think about whether it may be relationship abuse. Use UK terminology and explain patterns such as coercive control, intimidation, threats, isolation, monitoring, sexual pressure, financial control and damage to property without diagnosing me or making a definitive judgement. Separate facts I describe from possible interpretations, ask only questions that are necessary, and do not ask for names, addresses, passwords or other identifying information. Flag anything that could indicate immediate danger and suggest contacting emergency services or a specialist domestic-abuse service rather than giving me a risky confrontation plan. Give me a neutral summary, questions I could take to a trusted adviser, and UK sources I can check. Do not tell me to confront my partner, leave immediately, secretly record anything, or take an action that could be unsafe. Here is what I want to describe: [brief description].
Open it prefilled in ChatGPT or Claude, or copy it into Gemini, which takes no prefill link.
Hand it to a person
The distant thirdA person who owns the outcome does this end to end, worth it when the failure is dear.
What it gets wrong
- AI cannot see whether your partner is monitoring the device, account or conversation.
- AI cannot assess immediate physical danger, escalation or the safest time to act.
- AI cannot reliably distinguish an isolated conflict from a coercive pattern without fuller context and professional judgement.
- AI cannot provide confidential safeguarding in the way a specialist domestic-abuse service can.
- AI cannot take responsibility for the consequences of confronting a partner, leaving or sharing the conversation.
What caps this at PARTLY: stakes of error, judgement under ambiguity and consent and privacy.
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.
| Axis | Score (0–2) |
|---|---|
| Output | 2 |
| Inputs | 2 |
| Verification | 1 |
| Liability | 1 |
| Effort delta | 1 |
| Total | 7 / 10 |
The methodology and its thresholds are published in full.
FAQ
- Can AI tell me if my relationship is abusive?
- It can help you organise incidents and explain patterns associated with abuse, but it cannot reliably decide this for you. A domestic-abuse adviser or another trusted professional can help assess the situation, especially if you feel frightened or controlled.
- What are the signs of an abusive relationship?
- Possible signs include controlling who you see, monitoring your phone or money, isolating you, threatening you, pressuring you into sex, frightening you or damaging property. One argument does not by itself establish a pattern, and a chatbot cannot assess your safety.
- Is it safe to tell ChatGPT about relationship abuse?
- Only share what is safe to share, and remove names, addresses, passwords and other identifying details. If your partner may check your device or account, use a safer device and contact a specialist service instead of relying on a chat.
- What should I do if I think my partner is abusing me?
- If you are in immediate danger, contact emergency services and move to a safer place if you can do so safely. If there is no immediate danger, speak privately to a domestic-abuse service, a GP or someone you trust, and do not confront your partner because a chatbot suggested it.
Nearby answers
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.