Most people have had at least one experience of saying too much. The conversation that tipped past what felt appropriate, the emotional disclosure that made the other person go quiet in a way that was clearly uncomfortable, the overshare that you replayed on loop for the next three days. The lesson tends to stick, and it tends to produce the opposite problem: the systematic undersharing that follows, the holding back, the careful rationing of self that protects against the vulnerability hangover but also keeps everything at a surface level you did not actually choose. Finding the right amount to share with another person is genuinely difficult. It involves reading the room, knowing the relationship, having a sense of what the moment can hold. These are skills, and like most skills, they improve with practice. But most people do not have many safe practice opportunities for emotional disclosure. The stakes are high, the contexts are varied, and the feedback is rarely clear.

Why the Balance Is So Hard to Find

The oversharing/undersharing problem is partly a calibration problem — you are trying to match your disclosure to what the relationship and moment can absorb. But it is also a self-knowledge problem. Many people do not fully know what they want to say until they start saying it, and by the time they realize they have gone further than intended, the disclosure has already happened. The reverse is also true: people who are practiced at restraint often do not realize how much they are holding back until they notice a relationship feels thin in ways they cannot account for. Research from the Journal of Personality and Social Psychology has found that errors of disclosure tend to go in one direction over time: people initially overshare in new relationships and then overcorrect toward undersharing once they have experienced the discomfort of revealing too much. The overcorrection is often more damaging to intimacy than the original overshare, but it feels safer.

Using AI as a First Pass

One underexplored use of AI conversation is as a space for drafting disclosure before bringing it into a human relationship. Not scripting — that tends to produce a stilted, performed quality in real conversations — but processing. Working out what you actually think and feel about something, how much of it you want to share, what the core of it is versus the elaboration. When you have had the conversation once in a lower-stakes context, you often show up to the human conversation with more clarity about what you are actually trying to say. This is not a new idea in therapeutic contexts. Therapists often encourage clients to rehearse difficult conversations, not to predict the exact exchange but to develop a cleaner sense of their own position before entering a context where the position will be challenged or complicated by the other person's reactions.

The Practice Dimension

There is also a straightforwardly practice-based benefit. People who struggle with disclosure — in either direction — often benefit from more experience with the act itself. AI conversation provides that experience without the social cost of getting it wrong in a relationship that matters. You can be more open than you would normally be and observe what that feels like. You can notice where the self-editing kicks in and why. You can experiment with saying the thing you usually do not say and see what happens to the thought when it is spoken.

The Tangent: Why We Conflate Vulnerability With Weakness

There is a cultural inheritance that treats emotional disclosure as a kind of risk exposure — the more you reveal, the more leverage you have handed to the other person. This framing is not entirely wrong, but it is too simple. The research on vulnerability and relationship quality, much of it associated with Brené Brown's work at the University of Houston, consistently finds that the willingness to disclose — to be seen imperfectly — is a prerequisite for genuine closeness, not a naive relinquishment of power. The protection achieved by not disclosing is real, but so is the cost.

Getting Better at the Balance

Getting better at the oversharing/undersharing balance is a real goal, and it is achievable with practice. AI as a low-stakes environment for practicing disclosure is one tool among several. The point is not to perfect the calibration so you never make an uncomfortable disclosure again. The point is to become fluent enough in your own inner life that what you offer others is chosen rather than defaulted to.