In September 2025, photographers who had registered for the authenticity service on the Nikon Z6 III got an email. Part II of this series covered the flaw that forced it. This one starts with the email itself. It said a technical problem had been found. It said the digital certificates loaded onto their cameras since the feature launched would be invalidated, and that the credentials already attached to their photos were, in the email’s words, “no longer valid and cannot be used as proof of provenance.” It apologized. It promised an announcement when the service came back. It gave no address to write to and described no way to ask a question.
Nikon did the right thing by revoking. The flaw was real and the credentials could not be trusted. What interests me is the part the email did not contain. Across three essays this series has watched systems take something from a photographer and say nothing about how to question it. Research on other systems says that being told why, and being heard, can change how people take a decision they cannot reverse, even when it changes nothing else. I can find no published measurement of that in a provenance system. So this piece lays out what that research supports, where it stops, and what it would take to find out here.
The question, and where the answers come from
Part I set out a hypothesis and said plainly that it was untested. Authority exercised without an adequate account of how it can be challenged draws down acceptance over time. Part II found that a dispute crossing the six organizations in a Content Credentials chain has nowhere to go, and Part III found that the photographer is named in the data and a party to none of the governance. Neither tested the hypothesis. This is the installment that owes it evidence.
The evidence I have found is old, it comes from other systems, and most of it was gathered for other reasons. A 1990 laboratory experiment on setting work goals. Surveys and server logs from Reddit. Facebook users rating hypothetical moderators. Twitter users after an enforcement action. None of it was designed with a photograph in mind. Some of it is experimental and some is association, and the difference matters at every step. What this body of work can do is establish that the concern is real somewhere. What it cannot do is establish that the concern is real here. I am borrowing old tests to examine a modern worry, and I will say so each time the borrowing shows.
The oldest and plainest version of the idea belongs to the psychologist Tom Tyler, who has spent decades studying why people accept decisions from authorities. His best-known finding comes from Chicago traffic court. Under the practice at the time, defendants who showed up in person had their cases dismissed. No fine, no record. Tyler interviewed them and found they were angry anyway. Winning turned out to be “less important to them than having their day in court.” They had received the best possible outcome through a process that never looked at them, and the process was what they remembered.
That is the question in one image. When a decision goes against you and cannot be changed, does it matter whether anyone told you why, or heard what you had to say? The provenance chain is now making decisions of exactly that shape about people’s work.
