Whenever I come across posts stating that their AI imagery is indistinguishable from "reality," or that their goal is to make imagery that is indistinguishable from "real photography" (an oxymoron), I always ask, "Why."

It's not a trick question, but I get either no response or blocked (so they think), especially if I point out that reality is more than making an image look like it's a photograph. IT'S ABOUT THE REALITY OF THE SUBJECT. I instantly recognized each of the examples shown as fake.

If anyone ever creates an image that looks like a photograph with details about the subject (such as a species name, common or taxonomic) and a location, date, and time of day of the image that passes my scrutiny, then we can revisit this.

The only way to game that test is to post a photograph with all the details I mentioned (maybe with the EXIF data usable as "prompts") and claim it was AI-generated. I don't put that past anyone who is either blind to the unreality of this crap, or simply doesn't care because they have a financial interest in perpetuating these falsehoods to those who don't know better.

The ominous statements and questions raised are good but miss the target issue. This tech can never depict reality, but for the mass public who don't know or care about how the tech works, how it is trained, what it is and is not capable of, it doesn't matter. They will never discern the real from the unreal. The grim reality, above and beyond arguments over the reality or utility of AI-generative "art" and lack of discernment, is that this unreality game is more like hand grenades than horseshoes, and that's a very real and dangerous problem.