It might be way out there, but the who situation makes me think of a piece by philosopher Slavoj Zizek. The piece is about a link between belief in existence of God, or some other transcendent Big Idea, and limitations.

If one replaces the Big Idea, with the concept that "reality is and should be reality", then it stands to reason that either there will be more explicit censorship and/or that more importance will be placed on reputation/status/interpersonal relationships, and beside these two other forms of verifyability.

On the flipside, if one looks at what the majority of people actually create with it, you get things which are mostly (")productive(").

In a sense it is similar to the invention of the printing press: with it people could relatively rapidly publish anything, but in the end reputation of good publishers is what covers the majority of the market, and although there are still people who self-publish how far their reach is depends on credentials (in a broad sense of reputation, status, relationships & accomplishments).

what I am more worried about than fake X, is the opposite consequence. Because it becomes easier to create fake X, it should also become more "reasonable" to claim X is actually fake. Since deepfake detection is practically an arms race with no end in sight a different solution than AI is needed for this and consequences depend a lot on what that looks like.