Reassurance Seeking and Anxiety: Why Reassurance Stops Working

What is reassurance seeking?

You send someone an important message. Several hours pass without a reply.

You ask a friend: "Do you think they're annoyed with me?"

Your friend says: "No. They're probably just busy."

You feel better.

Twenty minutes later, another thought appears: "But what if they're only saying that to make me feel better?"

So you ask again.

That second round tells us something important. You may no longer be looking for information. You may be looking for certainty.

Psychologists use the term `reassurance seeking` for behaviours intended to reduce doubt, perceived threat, or anxiety by obtaining confirmation from another source.

A 2011 study by Rector and colleagues examined reassurance seeking in 283 people receiving treatment for social anxiety, generalized anxiety disorder, panic disorder, or OCD. Three broad themes emerged: reassurance about decisions, reassurance about relationship security, and reassurance about perceived threats or anxiety.

Those categories still capture much of what reassurance seeking looks like in everyday life.

What can reassurance seeking look like?

It is not always as obvious as repeatedly asking someone whether everything is okay.

• "Are you sure you're not angry with me?"
• "Do you think I made the right decision?"
• "Does this symptom sound serious?"
• "Was that message rude?"
• "Do you think I embarrassed myself?"
• "Are you sure I didn't make a mistake?"
• Repeatedly checking whether someone is online or has read a message.
• Googling the same health symptom after already reading an answer.
• Re-reading an email to confirm that nothing sounds wrong.
• Comparing the same decision across multiple forums, reviews, or people.
• Mentally replaying an event to reassure yourself that you did not do something wrong.
• Asking an AI essentially the same question with slightly different wording until the answer feels certain enough.

Reassurance itself is not unhealthy

This distinction is essential.

Humans reassure each other constantly.

A child asks, "Will you be here when I wake up?" A partner asks, "Are we okay?" Someone receiving medical treatment asks whether a symptom is expected.

Those interactions can communicate care, provide useful information, and help relationships feel safe.

The problem is not reassurance.

The problem is when reassurance becomes repetitive, increasingly specific, or necessary every time uncertainty appears.

Researchers generally call this `excessive reassurance seeking` rather than treating all reassurance as pathological.

A 2019 study of 738 people receiving CBT found reassurance seeking across panic disorder, social anxiety disorder, generalized anxiety disorder, and OCD. Its content differed, but the behaviour was not confined to one diagnosis.

Why does reassurance feel so effective?

Because, in the moment, it often is.

Anxiety produces uncertainty and a sense that something needs to be resolved.

You ask: "Do you think everything is okay?"

Someone says yes.

The uncertainty temporarily narrows. Anxiety falls.

That relief is real.

A 2015 study by Salkovskis and Kobori examined naturally occurring reassurance episodes in people with OCD, panic disorder, and healthy controls. In the anxiety groups, reassurance was associated with short-term relief, followed later by a return of discomfort and an urge to seek reassurance again.

The study relied on retrospective self-report, so it cannot establish the exact mechanism in every situation. But it captures the paradox well: reassurance can genuinely help now while still leaving the original uncertainty available to return later.

The reassurance loop

A common reassurance cycle looks something like this:

1. 1Something uncertain happens.
2. 2Your mind generates a threatening interpretation.
3. 3Anxiety rises.
4. 4You ask, search, check, or seek confirmation.
5. 5You receive reassurance.
6. 6Anxiety decreases.
7. 7A new doubt appears.
8. 8You seek reassurance again.

Why does the doubt come back?

Because reassurance often answers the current question without removing uncertainty itself.

Someone tells you: "I'm sure your headache is nothing serious."

That may help until your mind produces:

"But how can they know?"

"What if my case is different?"

"What if I didn't describe the symptom properly?"

"What if it gets worse tomorrow?"

There is nearly always another possibility if the standard is complete certainty.

That is why reassurance seeking connects so closely with intolerance of uncertainty.

The real question underneath the reassurance may not be `What is most likely?`

It may be:

`Can somebody guarantee that the feared possibility is impossible?`

Very few important things in life come with that guarantee.

Information seeking vs certainty seeking

This is probably the most useful distinction in the entire topic.

Information seeking asks:

`What do I reasonably need to know?`

Certainty seeking asks:

`How can I make sure there is no possibility that I am wrong?`

The behaviours can look almost identical from the outside.

You may search online, ask a doctor, message someone, or request another opinion in either case.

The difference is what happens after you receive a reasonable answer.

• Information seeking: "I don't understand this medication side effect, so I'll ask my doctor."
• Certainty seeking: "The doctor said it's expected, but I'll ask three forums to make absolutely sure."
• Information seeking: "My manager's instruction is unclear, so I'll ask what they want."
• Certainty seeking: "They answered, but I'll reread their message repeatedly to make sure they're not secretly unhappy."
• Information seeking: "I'll compare three options before deciding."
• Certainty seeking: "I'll keep researching until I know which decision can never turn out badly."

Three questions that can help you tell the difference

1. 1Has anything materially changed since I last checked or asked?
2. 2Would another answer give me information I can actually use?
3. 3Or am I mainly hoping the next answer will make the uncertainty disappear?

Reassurance seeking in relationships

Relationships naturally contain uncertainty because you never have direct access to another person's thoughts.

That makes them fertile ground for reassurance loops.

"Are you upset with me?"

"Do you still love me?"

"Did I do something wrong?"

"Are you sure we're okay?"

There are moments when asking one of these questions is healthy communication.

If your partner has become unusually distant, asking whether something is wrong may be more constructive than guessing.

The pattern becomes different when the answer does not stick.

Your partner says: "We're okay."

Five minutes later:

"Are you sure?"

An hour later:

"You sounded different when you said it. Is something wrong?"

The conversation has moved away from resolving a relationship problem and towards regulating uncertainty.

Relationship reassurance can be more complicated than 'stop asking'

The interpersonal research is not simple enough to justify telling everyone to stop seeking reassurance.

Different attachment patterns, relationships, and circumstances change how reassurance functions.

A 2022 daily-diary study of 110 couples found associations between attachment style, trust, and daily excessive reassurance seeking, but the interpersonal effects differed depending on the partners involved.

This is one reason blanket advice can backfire.

A healthy relationship includes asking for comfort, clarification, affection, and support.

The useful question is whether the interaction creates genuine connection or whether both people are being pulled into repeatedly solving the same uncertainty.

Reassurance seeking and mind reading

Relationship reassurance often begins with an interpretation of information you cannot directly know.

A short reply becomes: "They're angry."

A quiet expression becomes: "They're disappointed in me."

A delayed message becomes: "They're losing interest."

That is where reassurance seeking and the mind reading cognitive distortion often meet.

You try to resolve the uncertainty by asking somebody else what the person must be thinking.

Sometimes the better response is direct communication.

Sometimes the answer genuinely is unavailable.

Reassurance seeking and catastrophizing

The more severe the predicted outcome, the stronger the urge to make sure it cannot happen.

A minor symptom becomes: "What if this is something serious?"

One mistake becomes: "What if I lose my job?"

An awkward conversation becomes: "What if I've destroyed the relationship?"

This makes reassurance a natural partner to catastrophizing.

The catastrophe creates urgency.

Reassurance temporarily removes it.

But if the underlying belief remains `I need to know this cannot happen`, another version of the question can easily return.

Health reassurance and Googling symptoms

Health is an especially difficult area because getting accurate medical information and appropriate medical care is important.

The answer is not to ignore symptoms or avoid doctors.

The relevant pattern is repeated searching or reassurance after appropriate information has already been obtained, particularly when each search generates new possibilities and more anxiety.

A 2019 systematic review and meta-analysis examined 20 studies involving 7,373 participants. Health anxiety was positively associated both with online health-information seeking and with cyberchondria — repeated online health searching associated with distress.

That association does not mean researching health information causes health anxiety. People who are already worried about their health are naturally more likely to search.

But online information is uniquely capable of providing an endless supply of new diagnoses, edge cases, anecdotes, and reasons to check again.

Medical reassurance still has an important place

Do not use the idea of reassurance seeking to dismiss genuine medical concerns.

New, severe, persistent, worsening, or unexplained symptoms can require appropriate assessment.

Following medical advice, completing recommended investigations, and asking questions needed to understand your care are not psychological failures.

The distinction is between appropriate healthcare and using repeated medical checking primarily to obtain a feeling of certainty that medicine often cannot provide.

If you are unsure which category you are in, that itself can be something to discuss openly with an appropriate healthcare professional.

Online reassurance can feel easier than asking another person

The internet has changed reassurance seeking because you no longer need another person to participate.

Search engines, forums, social media, review sites, and discussion communities can all provide reassurance privately and immediately.

A 2025 study by Parsons and colleagues compared motivations for interpersonal and online reassurance seeking in people with and without OCD.

Participants described emotional support and personalization as reasons for seeking reassurance from people. Online reassurance was often attractive when interpersonal concerns made disclosure more difficult, and participants with OCD were more likely to mention concerns about the personal consequences of revealing their worries to others.

The study was relatively small and specific to OCD, so it should not be generalized to every anxious person.

But it highlights an important feature of online reassurance: it can remove some of the social friction that normally limits how often you ask.

Can AI become reassurance seeking?

Yes — depending on how it is used.

Suppose you ask an AI: "Do you think my friend is angry with me?"

The AI gives a thoughtful answer and reminds you that there are other possibilities.

That may be useful reflection.

But then you ask:

"But based on this exact message, do you think they're angry?"

Then:

"What percentage chance would you give it?"

Then:

"Are you sure I'm not missing a sign?"

At that point, the tool may no longer be helping you examine the thought. It may be functioning as an endlessly available source of reassurance.

Why AI can make the reassurance loop unusually easy

General-purpose AI has several characteristics traditional reassurance sources do not.

It is available at any hour. It does not become visibly frustrated. You can rephrase the same question indefinitely. It can generate highly specific responses to every new doubt.

A 2026 perspective by Golden and Aboujaoude proposed a CBT-informed model in which general-purpose AI chatbots can become a vehicle for reassurance seeking, checking, confessing, perfecting, and other avoidance behaviours in OCD and anxiety.

The authors describe this as an emerging conceptual framework rather than settled causal evidence. Direct empirical research on AI-driven reassurance loops is still limited.

The Anxiety and Depression Association of America now similarly cautions that repeatedly asking AI to confirm safety around anxiety or OCD fears can function as reassurance seeking.

That does not mean AI mental-health tools are inherently harmful

The broader evidence is more nuanced.

A 2026 systematic review and meta-analysis examined randomized trials of mental-health chatbots. Across 34 studies involving 7,621 participants in the anxiety analysis, chatbot interventions produced a small average reduction in anxiety symptoms compared with controls.

That literature includes structured mental-health interventions and should not be treated as equivalent to unlimited reassurance-seeking conversations with a general-purpose chatbot.

The useful distinction is again about function.

Is the tool helping you observe a thought, practise a skill, make a reasonable decision, or prepare for a conversation?

Or are you asking it to provide certainty again and again?

Reflection and reassurance are not the same thing

Consider these two conversations.

Reassurance: "Tell me that they're definitely not angry with me."

Reflection: "I'm interpreting their short message as anger. What evidence do I actually have?"

The first tries to close uncertainty.

The second examines the interpretation while allowing uncertainty to remain.

That is a central difference between an AI tool that becomes a reassurance machine and one that supports cognitive reflection.

How do you know if reassurance has become a loop?

There is no numerical rule for how many times you are allowed to ask.

Look instead at the pattern.

• You ask essentially the same question repeatedly.
• The relief lasts only briefly.
• Each answer generates a new exception or `what if?`.
• You keep changing the wording in search of a more convincing answer.
• You consult several people or sources after already receiving a reasonable answer.
• You feel unable to move on until somebody confirms that you are safe, correct, liked, healthy, or making the right choice.
• You increasingly distrust your own judgment after reassurance rather than before it.
• Checking or asking is taking significant time or interfering with relationships, work, sleep, or daily life.

What can you do instead of immediately seeking reassurance?

The goal is not to ban reassurance.

It is to create enough space to notice what you are actually asking for.

1. 1Write the exact question you want answered.
2. 2Ask whether genuinely new information is available.
3. 3Identify what you already know.
4. 4Notice whether you are looking for information or certainty.
5. 5If information is needed, choose one appropriate source.
6. 6If the question has already been answered, consider delaying the next check or reassurance request.
7. 7Decide what action is reasonable even if some uncertainty remains.

1. Write down the reassurance question

The urge can feel vague until you put it into words.

"I need to know whether she is annoyed."

"I need to know whether this symptom is harmless."

"I need someone to tell me that I made the right decision."

Now ask:

`Is this information available?`

Sometimes it is.

Sometimes only another person can answer it.

Sometimes nobody can.

2. Ask what you already know

Before seeking another answer, take inventory.

What did the doctor actually say?

What did your partner actually tell you?

What evidence did you already consider?

This is not an exercise in convincing yourself that everything is fine.

It prevents anxiety from acting as though no answer has ever been provided simply because the feeling of uncertainty has returned.

3. Decide whether anything has changed

This question can separate a new problem from an old doubt.

If your symptom has materially changed, new medical information may be appropriate.

If your partner has done something new that concerns you, a conversation may be warranted.

If nothing has changed except that anxiety has returned, another round of reassurance may be serving the feeling rather than the situation.

4. Delay the reassurance instead of banning it

For many people, `never ask again` is an unnecessarily harsh rule.

Try creating a small gap.

"I'll wait ten minutes before checking."

Then perhaps twenty.

The aim is not to prove that you can tolerate enormous distress.

It is to discover that the urge itself can change without always being answered immediately.

This can function like a small behavioral experiment.

Prediction: "If I don't ask, my anxiety will keep increasing and I won't be able to focus."

Then observe what actually happens.

5. Ask for support instead of certainty

This can be one of the most useful substitutions.

Reassurance request: "Promise me the test will come back normal."

Support request: "I'm really anxious while I wait. Can you stay with me for a bit?"

Reassurance request: "Tell me I definitely didn't embarrass myself."

Support request: "I'm stuck replaying the conversation. Can I talk about how uncomfortable that feels without trying to work out exactly what everyone thought?"

The difference is subtle but important.

Support helps you carry the uncertainty.

Reassurance tries to remove it.

Emotional support may be a genuine alternative

There is emerging experimental evidence for this distinction, at least in OCD.

A 2025 study by Causier and Salkovskis asked 36 participants with OCD to imagine receiving reassurance versus emotional support in personalized scenarios.

The imagined emotional-support condition produced a lower anticipated urge to seek further reassurance and was rated as more acceptable and helpful for managing emotions.

This was a small imaginal experiment, so it is preliminary rather than definitive.

But it supports an important clinical idea: reducing reassurance does not have to mean withdrawing warmth, connection, or care.

What if you are the person being asked for reassurance?

Repeated reassurance can be difficult for both people.

You may feel trapped between two bad options: keep answering the question indefinitely or refuse and seem uncaring.

There is a middle ground.

You can validate the emotion without pretending to possess certainty you do not have.

• "I can see this is making you really anxious."
• "We've already gone through what we know. I don't think another answer from me will make you more certain."
• "I can't promise what will happen, but I can stay with you while you wait."
• "Do you want help deciding what you can actually do next?"

Do not turn reassurance reduction into punishment

This is particularly important in clinical conditions such as OCD.

Suddenly refusing all reassurance without explanation can be experienced as rejection and may not address what the person actually needs.

The 2015 reassurance study itself cautioned against simplistic advice to loved ones to simply stop providing reassurance.

For OCD, severe health anxiety, or other significant anxiety conditions, reassurance patterns are best addressed within an appropriate treatment plan rather than through improvised confrontation.

What does CBT do with reassurance seeking?

CBT generally looks at what the behaviour is doing in the anxiety cycle.

Does reassurance provide short-term relief?

Does it prevent the person from learning that uncertainty can be tolerated?

Does it function like checking or another safety behaviour?

Treatment may involve examining beliefs about threat and uncertainty, reducing repetitive checking or reassurance, approaching feared situations, testing predictions, and learning other ways to respond to distress.

The 2019 study of 738 CBT patients found that reassurance seeking decreased over treatment across panic disorder, social anxiety, GAD, and OCD, and reductions were significantly associated with improvements in disorder-specific symptoms.

That finding is correlational within treatment. It does not prove that reducing reassurance alone caused the improvement.

But it supports reassurance seeking as a clinically meaningful behaviour rather than a trivial habit.

What does the broader research say?

Reassurance seeking has been studied far less extensively than anxiety itself, so claims should remain proportionate to the evidence.

The 2011 Rector study provided evidence for reassurance seeking around decisions, attachment, and threat across multiple anxiety disorders.

A 2012 study by Cougle and colleagues found that greater excessive reassurance seeking was associated with social-anxiety, GAD, and OCD symptoms even after accounting for several other psychological variables. Some forms of reassurance seeking also predicted changes in social-anxiety and GAD symptoms one month later.

The 2019 clinical study then showed reassurance seeking changing alongside CBT outcomes in a much larger treatment sample.

More recently, a 2023 ecological momentary assessment study followed 104 young adults for 14 days. On days when people reported relatively greater anxiety, they also tended to report greater reassurance seeking. Intolerance of uncertainty was associated with reassurance seeking at baseline.

And a 2026 study of 204 people with high levels of worry found that intolerance of uncertainty and cognitive avoidance independently predicted reassurance seeking and other worry-related behaviours.

Taken together, the research supports reassurance seeking as a transdiagnostic anxiety-related behaviour, but it does not justify treating every request for reassurance as maladaptive.

How Reframer should respond to reassurance seeking

An AI reflection tool has to be careful here.

If every anxious question receives a confident guarantee, the tool can accidentally become part of the same reassurance loop it is supposed to help untangle.

If you enter: "Do you think my partner is angry with me?" the most useful response may not be: "No, I'm sure they're not."

A more useful process is to separate:

• What actually happened.
• What you are interpreting.
• What evidence you have.
• What remains unknown.
• Whether there is a real action you can take.
• Whether you are asking the same uncertainty to be solved again.

How Reframer approaches the loop

Reframer is designed around reflection rather than certainty.

If the thought contains an assumption, cognitive reframing can help you examine it.

If you are trying to infer another person's private thoughts, recognising mind reading may be useful.

If one frightening possibility has taken over, the issue may be catastrophizing.

If there is a practical problem, you may need a plan rather than another interpretation.

And if the honest answer is `I don't know yet`, working with intolerance of uncertainty may be more useful than generating another reassuring answer.

A simple test: information, support, or certainty?

The next time you feel the urge to ask, search, or check, try identifying which of three things you actually need.

1. 1Information — there is something relevant I genuinely do not know and can reasonably find out.
2. 2Support — I know what the situation is, but it feels difficult and I do not want to carry it alone.
3. 3Certainty — I already have a reasonable answer, but I want somebody to guarantee that the feared possibility cannot happen.

Sometimes reassurance is the right response

An article about excessive reassurance can easily swing too far in the other direction.

Sometimes someone genuinely needs an answer.

Sometimes a child needs to hear that a parent is coming back.

Sometimes a partner needs explicit affection.

Sometimes a patient needs a doctor to explain what a test result means.

Sometimes a colleague needs to know whether their work meets expectations.

Psychological flexibility is not refusing reassurance on principle.

It is being able to notice when the answer is providing useful information or connection — and when you are asking the same answer to do the impossible job of eliminating uncertainty altogether.

Reassurance asks for certainty. Reflection asks what you can do with uncertainty.

There is nothing weak or irrational about wanting reassurance.

It makes sense to reach for something that reliably makes anxiety quieter.

The difficulty is that short-term relief and long-term resolution are not always the same thing.

Sometimes you need an answer.

Sometimes you need another person's support.

Sometimes you need to act.

And sometimes the question has already been answered as far as it can be.

At that point, asking again may not bring you closer to certainty.

The more useful question becomes:

"What do I actually know — and can I take the next step without knowing the rest?"