To expand on this point further, the phenomenon of synthesizing accounts of a person saying something has already existed for thousands of years.
If I tell you that "Trump said that low income families will receive unconditional, federally sponsored medical coverage," you will not immediately take my word on it. You will critically analyze my statement by scritinizing who I am to make this claim, and the claims congruency with what you already know to be reality. I am not a reputable publication, so you are not likely to believe me, because the claim is contrary to what is already known.
The same will become true of video. A video of what looks like somebody saying something will become as useful as hearsay testimony. We will, through necessity, begin critically scrutinizing the source of footage, not just taking it's content at face value.
In many ways this already exists in the photography world. Convincing photo manipulation is old technology at this point, and I think we've adequately adapted to it.