Giving an AI assistant a person's name and job title seems to earn trust, but it only borrows it. Whoever interacts with an AI must know it: it is the most cross-cutting obligation of the European AI Act, in force since August 2026, which also requires synthetic content to be labeled as such. Besides an obligation, it is a trust policy: a user who discovers that "their advisor" was a machine does not forgive the concealment, but does forgive the honesty. That is why a product's transparency duties are decided before building, as part of the Product Owner's work: the assistant presents itself as what it is from the very first message, because trust is earned with the truth.
Want to keep pulling the thread? This topic unfolds in Product Owner (v 2.0) > Products that incorporate AI > Governance and regulation > The stable frameworks.
We will send you every new strip when it is published.