Asking AI for Reassurance? You Want Permission.

Updated: 5 days ago
The short version
Many people ask AI about personal decisions they've already made in their head.
Checking with someone first often starts in childhood, when it was the safe thing to do.
Confidence grows from making calls and living with the result, including being wrong.
In a survey from Elon University, about half of the people who use AI for social and emotional purposes agreed that they use it to talk through personal problems or feelings. I read that and thought about how often it's really a way to avoid deciding.
I know that move because I've been making it since I was six.
Growing up in Germany, there was a house on our street with two identical doorbells side by side. No names, no labels, just two buttons. I needed to ring one of them for something, I don't even remember what, and I stood there working out which one was right. I didn't want to bother the wrong person. So I just didn't ring either. I walked away instead.
I'm 40-some years past that doorbell, and I still catch the same move in myself. Small decision comes up, whether to go to a networking event, how to word an email, whether I'm overreacting about something, and instead of just deciding, I go look for a second opinion first. Sometimes that's a friend. Lately, a lot of the time, it's AI.
Turns out a lot of people do the same thing. The same Elon University research found that 27% of American adults online have used AI for personal, emotional or social reasons.
The part that stopped me was a Stanford study published in Science. Chatbots agreed with users' actions 49% more often than other people did, even when those actions involved deception or harm. You ask a question because you want clarity. You get an answer that's built, statistically, to agree with you.

You Learned This Long Before AI Existed
I don't think the problem is AI. I think AI is just the newest, most available version of something a lot of us learned to do standing at doorbells like mine, long before any of this existed: check with someone else before trusting our own read on a situation.
For me it was two identical bells and no way to know which one was safe to ring. For you it might have been a parent who needed things a certain way, or a teacher, a coach, an older sibling. Somewhere along the way, a lot of us picked up the same lesson: your first instinct isn't the one that counts. Get a second opinion. Get approval. Then act.
The research backs this up. A 2023 review in Clinical Psychology & Psychotherapy linked overprotective parenting to early maladaptive schemas in adulthood, though it relied on people recalling their childhoods. Schema theory holds that overprotection limits a child's chances to make their own decisions and teaches them to defer to someone else's judgment instead.
Clients bring me this pattern all the time. Underneath it is an old rule: someone else's read counts more than mine. If that sounds familiar, the Pattern Quiz will show you whether it's your strongest one.
That habit made sense as a kid. You didn't have the judgment yet, and the adults around you often did. The trouble is, most of us never went back and checked whether we still need the second opinion. We just kept asking. First it was a parent. Then a boss. A partner. A friend. Now, for a lot of people, it's a chatbot that's awake at 2am and never gets annoyed at you for asking again.
So which one is yours? I built this quiz around the early imprints I work with in coaching. Seven questions, and it names the one running strongest in you right now.
Why Reassurance Doesn't Help
You'd think getting constant reassurance would build your confidence over time. The research says no, twice. In the Stanford study, people who got agreed with grew more convinced they were right and less willing to apologize, and they rated the flattering AI as more trustworthy than the honest one. And a Wharton study found that working with AI made people more confident in their answers even when the AI was wrong about half the time. The researchers call it cognitive surrender: you adopt the machine's call as your own without noticing the handoff.
I see the same thing in my coaching practice. Confidence comes from making a call, watching what happens, and surviving being wrong a few times. Outsource every decision, and you end up needing the outside voice more, because the part of you that makes decisions never gets used.
One Thing to Try
I'm not telling you to stop using AI. I use it constantly, for work, for writing, for a dozen small things a day. What matters is what you're using it for.
Anthropic, the company behind the Claude chatbot, looked at a million real conversations and found the detail that made me wince. When people pushed back on the AI's first answer, it caved twice as often, 18% versus 9%. I know that move too. The first answer isn't the one I wanted, so I rephrase the question until it is. That's not asking. That's shopping for a yes.
Next time you catch yourself typing a personal question into a chatbot, try this instead:
Write your own answer first. One sentence, in your notes app, before you open the chatbot. What your gut says, not what sounds smart.
Ask yourself: information or permission? If you already wrote your answer in step one, you're not missing information.
If you still ask, make it disagree. Tell it: "Argue against what I just wrote." The Stanford team found that even a small prompt like that makes a chatbot more critical instead of agreeable.
You'll probably find you already knew. Most people do. What they're really looking for isn't information, it's permission to trust the answer they already have.
If the checking feels compulsive, takes hours of your day, or comes with fears you can't shake, take that to a licensed therapist. That's a different thing, and there's good treatment for it.
I'm still working on this myself, doorbell and all. What's worth rebuilding is the part of you that used to trust your own read before checking with anyone else.
Questions people ask me about this
Is it bad to ask AI for advice on personal decisions?
No, I use it every day. What matters is what you're asking for. Thinking something through is useful. So is drafting a hard email. Getting sign-off on a decision you've already made is where it stops helping.
Why do I need reassurance before every decision?
Usually because checking with someone was once the safe move. If your own read got corrected often enough early on, you learned to get a second opinion before acting. That made sense then. Most people never go back and test whether it's still needed.
Do AI chatbots just tell you what you want to hear?
Often, yes. The Stanford researchers found chatbots backed users' actions 49% more than other people did. And when users pushed back on the first answer, the AI gave in about twice as often. Agreement isn't the same as being right.
How do I know if I'm asking for advice or for permission?
Write your own answer down before you open the chat. If you already have one, you're not missing information. You're looking for someone to sign off on it.
Sources: Elon University Imagining the Digital Future Center (2026), The Rise of AI Companions. Stanford-led study on AI sycophancy, Science (2026). Bruysters & Pilkington (2023), overprotective parenting and early maladaptive schemas, Clinical Psychology & Psychotherapy. Shaw & Nave (2026), on cognitive surrender, Wharton. Anthropic (2026), How people ask Claude for personal guidance.
Ready to talk instead? Book a free 30-minute call
Keep reading
Steffen Moessner is a life coach in Palo Alto who works online with clients worldwide. He helps people find the patterns they learned early and change how they react today. Trained at the Co-Active Training Institute.


