What if a coach colleague told you this?
"I built an AI tool so my clients can reach me between sessions, if they need a thought partner in the moment. They know I can see the conversations. But a client recently revealed something dark and disturbing about themselves to the bot. The entry was at 3 a.m. I don't think that they would have disclosed this during a session."
In this hypothetical scenario, the client signed a consent form. But alone with a screen in the early hours of the morning, consent isn't what they're thinking about. Psychologists call this the "online disinhibition effect," where people share more deeply with a machine than they would face-to-face. The absence of a human presence creates a false sense of solitude.
Now this coach faces a choice. Bring it up, and the client realizes that what felt like a private moment wasn't really. Stay silent, and the coach carries knowledge their client doesn't know they possess. Either way, the dilemma will likely influence how the coach shows up in the next session, intentionally or not.
Disclosure doesn't resolve this. The consent was informed. The access was transparent. And the dilemma remains.
This scenario is not hypothetical for long. As AI tools are increasingly integrated for between-session support, the online disinhibition effect makes this a high-probability risk for modern coaches. There is a real psychological gap between signing a consent form and the emotional experience of a late-night disclosure. Clients may intellectually know a coach can see the logs, but at 3am, alone with a screen, that knowledge recedes. And when the disclosure implies risk of harm, the stakes sharpen further: stay silent and you may be failing in your duty of care; speak up and the client may feel watched rather than supported. AI can handle data. It cannot handle the relational consequences of that data.
← All articles