Introduction
The problem is not simply that people are untrustworthy. It's that the instrument you'd normally use to assess them has been through several years of heavy weather.
You have trusted people who used you. You have been used by people who trusted you. You have lied fluently to good people and been lied to by people you'd have vouched for. Somewhere in there the calibration went, and now every judgment about who is safe comes with a footnote questioning the judge.
Trust Is Not One Thing
The most useful move is to stop asking whether someone is trustworthy in general.
Trustworthy with what? Different people can be trusted with different things, and lumping them together produces either indiscriminate disclosure or total isolation.
Someone can be reliable in a crisis and incapable of keeping a secret. Someone can be discreet and useless to call at 3am. Someone can love you completely and be entirely unable to hear about your relapse without falling apart.
So the question becomes specific: can this person be trusted with this particular thing? That's answerable, from evidence you already have.
Evidence You Already Have
You don't need to guess. Look at what they've already done, with other people's material.
Have they told you someone else's secret? Then they will tell yours. This is close to a law. Discretion is a general trait and it does not make exceptions for you, however much they love you.
How did they respond the last time someone brought them something difficult? With curiosity, or with panic, or by making it about themselves?
Do they do the small things they say they'll do? Reliability in trivial matters is the best available predictor of reliability in serious ones, because it's tested more often.
Have they ever used something against you in an argument? A person who reaches for ammunition when cornered will reach for this.
Start With Something Small
Since the judgment is unreliable, don't stake everything on a single assessment.
Trust is built by testing, incrementally. Tell them something modestly private — not the worst of it, something real but survivable — and then watch. Does it come back to you from a third party? Do they raise it again at a bad moment? Did the relationship change in a way you didn't want?
This sounds calculating. It is calculating, and it's what people with intact judgment do automatically. You are doing deliberately what most people do without noticing, because your automatic version needs supervision for a while.
Beware Both Directions of Error
Your instrument is off, and it can be off either way.
Too trusting: disclosing to whoever is in front of you in a moment of need, because the relief of saying it out loud overwhelms the question of whether this is the person. The 3am confession to someone you barely know is a real pattern, and it has consequences.
Too closed: deciding that nobody can be trusted, which is unfalsifiable, self-protective, and precisely the belief that keeps a person isolated and using.
The second error is far more dangerous than the first, and it disguises itself much better, because it looks like wisdom acquired through hard experience.
Professionals Are Structurally Trustworthy
This is worth understanding, because it's not about their character.
A therapist, a doctor, or a counselor is bound by confidentiality rules — and in the US, records from federally assisted substance use disorder treatment programs carry specific additional protections under 42 CFR Part 2, which is stricter than general medical privacy law and generally requires your written consent before records are disclosed.1
The exceptions are real and worth knowing: court orders, imminent danger to yourself or others, and mandatory reporting of child abuse, with variations by state and provider type. It's entirely legitimate to ask a provider directly, in the first session, what they are and aren't required to report. Good ones answer without defensiveness.
The point is that you're not relying on judgment about a person. You're relying on a structure that exists precisely because people in your position could not otherwise risk being honest.
Groups Solve a Different Problem
Peer support offers something else again: people whose own exposure is symmetrical.
The person across from you holds material at least as damaging as yours. That symmetry does a lot of work — it removes the power asymmetry that makes disclosure frightening, and it means the room has a strong collective interest in discretion.
It's not a guarantee. People talk. But the structure is more protective than confiding in a friend who has nothing at stake.
If Someone Has Already Betrayed a Confidence
It happens, and the damage is specific.
Someone told. It came back to you from a third party, or it was used against you, or it arrived in a custody filing. Whatever the mechanism, a piece of your worst material is now in circulation and you did not put it there.
The consequence is rarely the practical fallout. It's the conclusion — I will never do that again — which is entirely reasonable, entirely understandable, and functionally a decision to be alone with this.
The correct response is narrower. That person cannot be trusted with that kind of information. That is the finding. It does not generalize to everyone, however much it feels like it should, and the generalization is the thing that costs you.
Distrusting Yourself Is Part of This
The unstated half of the question.
Much of the difficulty in knowing who to trust is that you're aware of having been, recently, someone who couldn't be trusted. And a person who has watched themselves deceive people they loved has good reason to doubt their read on everyone.
That heals the same way any trust does: through accumulated evidence, gathered slowly. In the meantime, borrow structure instead of relying on instinct. Confide where the confidentiality is enforced by something other than someone's good character — including your own.
The Bottom Line
Ask what someone can be trusted with, not whether they're trustworthy. Use evidence you already have, especially how they've handled other people's secrets. Test in small increments rather than staking everything on one read. Be aware that "nobody can be trusted" is the more dangerous of the two errors and the better disguised. And where your judgment can't be relied on, use structures — professionals and peers — that don't require it.
Sources
- 42 CFR Part 2 protections — US Department of Health and Human Services. Understanding Confidentiality of Substance Use Disorder Patient Records (42 CFR Part 2). View source ↗