Should I use AI to analyze my relationship?

A chatbot has no independent view of your relationship. It only has the messages and framing you give it. It can be too eager to reassure you, even when the full picture is more complicated.

In one study of 11 models, chatbots often backed whichever side of a conflict the user took. The researchers found that models protected a user’s self-image more than humans did. Read the study.

Other research finds that this tendency to agree with a user can be rewarded in the feedback used to train AI assistants. Read the study.

AI can help you slow down and look at a specific thread. It can help you notice a repeated message, a change in tone, or the moment you backed away from what you meant.

That is different from asking it to judge your partner, predict what they mean, or give you a final answer about the relationship.

Do not use a chatbot as a therapist, a diagnosis, or a referee. It cannot decide whether someone is toxic, tell you what attachment style you have, or make the call about whether to stay or leave.

Research on using language models in place of mental-health providers found serious limits, including harmful responses in sensitive situations. Read the study.

People also tend to act on personal advice from chatbots. In a UK trial, many participants said they followed the advice they received, without measurable wellbeing gains two to three weeks later. Read the study.

Movecheck is not a chat about your relationship. You upload one thread and it looks at one narrower question: what did you do in this conversation?

It names only moves it can point to in the messages you sent. It does not diagnose either person, score the relationship, or decide who is right. You can see the full library of 28 moves, a sample reveal, or exactly how screenshots are handled.