For most of human history, it was very hard to determine whether what someone told us was true. Should we believe them? The answer came down to several factors: does the claim make sense with what we already believed to be true? Was the person who told us reliable? Were they truthful in the past? Were they gullible or skeptical themselves? Could anyone else confirm what they claimed? What did the evidence look like and could we examine it?

It was easier to vet the claims of something that happened in the recent past. It was very hard to vet the claim of something that just happened, particularly if it happen far away. For that reason, rumors were rampant in the old days. Someone they trusted told them something they heard from someone they trusted and they were now telling you. In fact, before the age of printing, this chain of communication was primarily how most information was conveyed, and it was extremely hard to weed out what was true and what was exaggerated or false.

The invention of photography changed this dynamic. We came to believe a photograph as evidence of truth. You might claim something, and I might not believe it, but if you showed me a photograph of it, I HAD to believe it. A photograph was inherently believable, unless obviously altered, in contrast to words, which were inherently malleable. When you viewed a photograph, it was innocent, unless proven guilty. Video had the same default. A video was inherently truthful, unless labeled otherwise.

The arrival of generative AI has flipped the polarity of truthfulness back to what it was in old times. Now when we see a photograph we assume it is fake, unless proven otherwise. When we see video, we assume it has been altered, generated, special effected, unless claimed otherwise. The new default for all images, including photographic ones, is that they are fiction – unless they expressly claim to be real. 

This claim for veracity can come in several ways. Increasingly, the origins of an image will be embedded in its metadata. It will code for its origins either as a generated image, or an unaltered image from a trusted camera. Secondly, an image can claim its source. Is the person or institution who provides the image, trustworthy and reliable in the past?

We will come to see that our default of “trust first and check later” was only a short temporary anomaly in our long history. We are back to the state we have been in for most of our time as humans, where we “check first and trust later.” The trust flip happened. Just recently, our initial default response to a photograph was to believe it as real, unless proven a fake. Now our default response to an image is to disbelieve as a fake, unless proven to be real.