‘Trust but verify’ is a translation of a Russian proverb made famous by former US President Ronald Reagan. In their paper, Graham et al appear to take an alternate view that might be summarised as trust or verify. The contrast highlights a general question: how do we come to trust in authorities? More specifically, Graham et al claim: (1) that UK Trusted Research Environments (TREs) are misnamed as future custodians for big health data because their promised verification systems actually negate the uncertainty that trust requires; (2) the public is mistaken if it believes such verification enhances trust; (3) the notion of building public trust in TREs is unclear or misconceived. In response, I propose a more relational, perhaps less reductionist account. I argue (1) that verification is itself a source of uncertainty, so it can’t extinguish the uncertainty needed for trust; (2) it’s nevertheless possible for verification to enhance feelings of trust thereby reducing our needs for the same; (3) trust is also social, even political, meaning institutions like TREs may become too big to fail—and end up shielding their ‘trusted’ brand by being less candid about inevitable flaws in their verification systems.