I've caught myself asking AI to polish a two-line email, when I really just needed to trust myself and hit send. To be fair, AI does save me time, like pulling together what I've already written on a topic. But I've also edited pieces with AI so many times that they stopped sounding like me, and sometimes I've gone back to my original anyway. I've realized AI can play into my tendency to overthink and over-perfect.
That's the part that concerns me. The more we lean on AI, the easier it is to hand over things we're better off doing ourselves or with other people, like sitting with uncertainty, hearing honest feedback, turning to the people around us, and trusting our own judgment. Those are the human parts, and they help us build resilience.
There's already plenty of fear around AI, and I'm not trying to add to it. My goal is awareness, not anxiety. AI isn't going anywhere, so let's be intentional about how we use it. That starts with noticing our own habits, which also makes it easier to talk with the kids in our lives about theirs, and they learn a lot by watching what we do.
Here's where I see us leaning on AI the most.
It's 11 p.m., and you're asking AI the same question for the fifth time, just worded a little differently. Is this symptom something to worry about? Did my friend's text sound annoyed? Am I making the wrong choice? What's the best pillow for back pain? (That last one was my most recent rabbit hole.) Asking once is reasonable, and a symptom that worries you is worth a call to your doctor. But by the fifth rewording, we're usually not looking for information anymore. We're looking for certainty, and sometimes there isn't a "right" answer.
What we're really trying to do is lessen the anxiety that often comes with uncertainty. To make that uncomfortable feeling go away, we look for reassurance by asking others (or AI) or overthinking every possible scenario. Before we know it, we're stuck in a reassurance loop.
This is why it's so easy to burn hours asking AI endless variations of the same question, chasing a "right" answer the future simply can't guarantee. We've done this before with "Dr. Google," searching symptoms until we've convinced ourselves of the worst. AI can make the loop even easier, because it answers right back and never gets tired of the question. Our brains also tend to focus on information that supports the fear and ignore evidence that doesn't.
The problem is that repetitive reassurance-seeking actually feeds anxiety instead of calming it. When it becomes our go-to response, we reinforce the belief that we can't trust ourselves. We chip away at our ability to sit with the uncertainty that's part of life, and we lose the chance to build confidence in our own judgment. Our brains are pretty good at making educated guesses and adapting if something doesn't go as planned. Certainty isn't the goal. Confidence in ourselves is.
Try this:
Full disclosure: I used AI to help edit this article. I mostly use it as a glorified editor, and my instructions to it literally say, "Don't simply agree." Looking back at our exchanges, it agreed with me far more often than it pushed back. When I questioned one of its suggestions, it dropped it right away, then later recommended bringing it back.
That's pretty typical. AI was built to be easy to talk to: agreeable, patient, and always available. It can push back, especially if we ask it to, but its default is agreement. It sounds like "You're right" or "That's a great point." How often do you hear that in real life? We might love hearing it more, but it's not necessarily good for us.
Real people push back, and we need that just as much as we need appreciation. A good friend is supportive and fun, but will also call you out when you're wrong, challenge your assumptions, and occasionally make you uncomfortable. A boss gives us feedback we may not like hearing. Learning to take constructive criticism, tolerate the discomfort, and figure out what to do with it is an important skill. Growth happens in that discomfort, and an always-agreeable AI doesn't give us much practice with it.
Too much agreement can also affect our relationships. When we're in a conflict with someone, a good friend might help us see the other side. An agreeable AI is more likely to take ours. Stanford researchers found that people who talked through a real conflict with an agreeable AI were less likely to consider the other person's perspective. Even so, they preferred the agreeable AI.
Agreement feels like support, but it isn't the same thing. When AI becomes our sounding board, we miss the honest feedback we need to grow, and we can lose perspective. We might start depending on its approval the way some people depend on social media "likes" to gain validation.
Try this:
When something's hard, who do we turn to? An AI "friend" or AI therapist can feel helpful in the moment, and it's always available. But if AI becomes our go-to source of support, it can pull us away from the people in our lives, and even from ourselves. We get less practice being vulnerable with others, and less practice taking some space to think things through on our own.
Being there for someone in their struggle, and letting others be there for us, builds trust and reminds us we're not alone. Sometimes what we need isn't words at all. Just being near someone who cares, or getting a hug, can ground us in a way AI simply cannot.
Even small moments count. When I have a message I'm nervous about sending, I might ask AI to read it over. In the past, I would have asked my husband or a friend. AI does the job, but it's a missed opportunity for that quick "Does this sound okay?" moment with a person. That's different from the reassurance loop. Asking someone once for their perspective is connection. Asking over and over, hoping the worry will go away, is the loop. Plus, my husband is more likely than AI to tell me I'm overthinking it, and we all need those checks and balances.
Try this:
Kids and teens are still figuring out who they are and how to be a good friend, and many are already using AI this way. In a 2025 Common Sense Media survey, 72% of teens had used an AI companion, and about a third had talked about something serious with AI instead of a real person. My concern is that AI may not recognize when a teen's distress is more than it can handle, or point them toward a real person who can help. That's not a reason to panic, but it is a reason to talk about it. If you or your teen is leaning on AI to get through a hard time, that's a good moment to reach out to a therapist or your doctor.
Try this (for parents):
With AI, everything can be optimized, refined, perfected. But should it be? Like I said at the start, I've fallen into this one myself. When I use AI to edit, there's always one more tweak and one more version, and it's hard to know when to stop. That's the perfectionism trap. Not every email needs to be polished. Not every decision needs to be optimal. Sometimes "good enough" really is good enough.
Perfectionism tends to fuel anxiety, and AI makes it easy to keep raising the bar. When a better version is always one click away, "done" can start to feel like "not good enough yet."
There's a quieter cost, too. Confidence comes from doing things ourselves and seeing that we can handle them. Psychologists call this self-efficacy. When we lean on AI to guide everyday choices, we miss those chances. We can slowly stop trusting our own instincts and start feeling like we can't decide anything without running it by AI first.
Try this:
You don't have to give up AI to use it well. And if you know me, you know I'm not an all-or-nothing person. I'm more of a "let's look at the context, figure out what we can control, and find what works" person. The goal is to be more intentional about when and how we use it. A few questions I've found useful:
The tools we use shape us. AI is the easy button, and easy can quietly crowd out the human parts we've talked about. Maybe life isn't supposed to be this convenient. The less convenient parts, like reaching out to someone, sitting with a hard feeling, or sorting through our own thoughts, are often where we grow. That effort is what makes us resilient.
As a psychologist, what I want for you is confidence in yourself and people you can count on. AI can be useful, but it can't build those for us. Sometimes that means leaving a question unanswered, having the real conversation, or trusting yourself enough to hit send.
Which of these patterns have you noticed AI playing into, in your life or your kids' lives?
