How to Stop Assuming the Worst About People
Updated 2026-09-04
The message says "ok." Not "ok!" and not "ok, sounds good," just the two letters and nothing after them. And somewhere in the next few seconds, without you choosing it, a whole story has assembled itself. They are annoyed with you. It was the thing you said on Saturday. They have been meaning to say something for a while.

By dinner you are being slightly cooler with them than usual, which they notice, and cannot explain, and now there is something between you that started as two letters on a screen.
Most of us have some version of this. Here is what is actually happening, and what helps.
Silence does not stay empty
The thing worth noticing first is that you did not decide to think badly of them. The story arrived already finished. That is not a character flaw and it is not something you are doing on purpose. Ambiguity is uncomfortable, and a mind handed an unexplained silence will reach for an explanation, quickly, from whatever is nearest.
What is nearest is usually you. The explanation your mind reaches for tends to be assembled out of what you already suspect about yourself. This is why the same short reply can leave one person entirely untroubled and leave another person certain they have ruined something. The message was identical. What it landed in was not.
That reframe matters, because it changes what you are dealing with. You are not being unusually perceptive about this person. You are watching your own mind do something ordinary and fast.
What the research actually says about this
There is a well known idea in psychology that we explain our own behavior by our circumstances and other people's behavior by their character. You were short with someone because you had a brutal day. They were short with you because that is who they are.
It is worth being careful with that idea, because it turns out to be less universal than it is usually presented. A meta-analysis of 173 published studies in Psychological Bulletin found that the classic actor observer asymmetry is far weaker than psychology long assumed, with average effect sizes close to zero. So the tidy version, that we always do this, does not hold up.
What the same analysis reported is more specific, and more useful. The asymmetry did hold under particular conditions, and two of them are worth sitting with: it held for negative events, and it held when the actor and observer were intimates. In other words, the pattern showed up most clearly in exactly the situation you are in when it hurts, something has gone wrong, and it involves someone close to you.
That is a meta-analysis of attribution studies, many of them using hypothetical scenarios, and it does not tell you anything about your particular person or your particular Saturday. What it offers is a reasonable prior. In the moment you are most confident you are reading someone correctly, you are also in the conditions where this bias is most likely to be operating.
Imagining their side is not the fix
The usual advice here is to put yourself in their shoes. It sounds right, and most people believe it works.
The evidence is not encouraging. Across 25 experiments reported in the Journal of Personality and Social Psychology, researchers tested whether being instructed to take another person's perspective made people more accurate about that person's thoughts and feelings, and found no consistent evidence that it did, with perspective taking if anything decreasing accuracy while occasionally increasing confidence. Their explanation was that imagining someone's perspective reduces self centered bias but does not supply better information in its place. You have swapped one guess for a more elaborate guess.
The confidence part is the part to watch. Sitting and thinking hard about why they sent "ok" does not get you closer to the truth. It gets you more sure of whatever you were already inclined to believe.
The same paper reported that getting a person's perspective directly, through conversation, did increase accuracy. Not imagining. Asking.
Ask the smaller question
Asking sounds like a bigger move than it is. You do not have to open with "are you upset with me," which puts the other person on trial and makes it hard to answer honestly even if the answer is no.
Ask something smaller and more ordinary:
- "Hey, is everything alright today?"
- "That read a bit flat, but I might be inventing that. All good?"
- "How was your day, honestly?"
Naming your own uncertainty, rather than the accusation, does most of the work. It gives them somewhere easy to land, including the very common landing of "oh, sorry, I was on the train."
There is also something worth knowing about questions generally. Across three studies of live conversations reported in the Journal of Personality and Social Psychology, researchers found that people who asked more questions, particularly follow up questions, were better liked by their conversation partners, an effect explained by being perceived as more responsive. Those were get to know you and speed dating conversations, not tense moments with someone you love, so do not stretch it too far. Still, the direction is a kind one. Asking is not an imposition on the relationship. It tends to be received as care.
When you cannot ask yet
Sometimes asking is not available. It is late, or they are travelling, or the thing is too tender to raise over text, and you have a night to get through with the story running.
Writing it down helps a little here, and it is worth being honest about how much. A meta-analysis of 146 randomized studies found that expressive disclosure had a positive and significant average effect, though a small one. Small is not nothing, particularly at eleven at night.
What seems to help most is not venting the story but separating it into parts. Write the two things apart from each other:
- What actually happened. They replied with one word. That is the whole of it.
- What I have added. They are annoyed. It is about Saturday. They have been meaning to say something.
Seeing those in two lists rather than one blended feeling is usually enough to loosen the grip. Not because the second list is definitely wrong, but because you can finally see that it is a list you wrote.
Using Murror for this
This is close to what Murror is built for. You open up to a caring AI that helps you understand what you are feeling and, just as importantly, helps you think clearly about the people you care about. On a night like the one above, you can say the whole spiral out loud without editing it, and be met with something that helps you tell the difference between what happened and what you have added, gently, without telling you that you are being ridiculous.
Over time Murror surfaces quiet insights about your relationships. If you keep arriving at the same worry about the same person, that is worth seeing, because it usually says something about the relationship that is easier to act on than to notice alone. Murror also suggests small, low pressure ways to show up for people, through Moments to Care and Connections, which in a situation like this is often just a nudge to send the ordinary question instead of spending another evening guessing.
And if something you work out privately turns out to be worth saying, you can turn that reflection into an optional takeaway to share with the person it concerns. Nothing is shared unless you choose to share it. Everything stays encrypted and private by default.
Murror is not therapy and it is not a substitute for the people in your life. It is a place to get clear enough to go back to them.
A gentler default
You will not stop your mind from filling silences. That part is fast and largely automatic, and trying to police it just adds a second thing to feel bad about.
What you can change is what happens next. The story arrives, and instead of treating it as information, you treat it as a first draft. Then you either ask a small question, or you write down the two lists and go to sleep.
That is a smaller change than deciding to trust people more. It is also the one that tends to work.
Frequently asked questions
Why do I always assume people are upset with me?
Because silence is genuinely ambiguous, and your mind will not leave it that way. When there is no information, something has to fill the space, and what fills it is usually built from whatever you already fear about yourself. If some part of you suspects you are too much, the short reply becomes proof of that. If some part of you suspects you are forgettable, the unanswered message becomes proof of that instead. It feels like reading the situation. It is closer to reading yourself, which is why two people can receive the same short text and arrive at completely different conclusions.
Is it not safer to expect the worst so I am not disappointed?
It feels safer, and that is worth respecting, because for a lot of people it was once genuinely protective. The trouble is that it does not actually work the way it promises. Bracing does not soften the moment if the bad thing happens, and it quietly costs you the hours in between, along with a certain warmth in how you speak to the person while you are waiting. You end up paying for the disappointment whether or not it ever arrives. Giving someone the benefit of the doubt is not naivety. It is declining to pay twice.
What if I give someone the benefit of the doubt and I turn out to be right about them?
Then you will know, and you will know it from something real rather than from a story you assembled while waiting. Benefit of the doubt is not a verdict and it does not commit you to anything. It is a decision about what to believe during the gap, before you have the information. If the pattern repeats and repeats, that is information too, and you are allowed to act on it. Most of the time it does not repeat, and you will have spent the gap being kind instead of building a case.
