People in general decide that images are either deepfakes or AI-Generated imagery unless proven otherwise. This change in sentiment has to happen before 2030.
Evidence for this claim will be taken from reputable newspapers. If newspapers like the New York Times, Financial Times, or Wall Street Journal assume that all images are AI generated or deepfakes unless there is a compelling reason to believe otherwise, then this claim resolves as YES.
We've had "deepfakes for text" (i.e. writing down a lie) for millenia and still we are not really at the point where "text is assumed to be made up unless proven otherwise".
Faking a screenshot of a tweet us trivial, yet people take them seriously.
So technology alone is unlikely to be enough to resolve YES. A non-trivial increase in people's willingness to lie (or assume others are lying) is required as well.
Note: a YES resolution might be easier to achieve if the market was limited to politics or other contentious areas, but it explicitly requires a general breakdown of trust even for say a photo of your family lunch on Facebook, which is IMHO a pretty high bar.