Technology is evolving but not in a vacuum, society's reactions also evolve in response. Today, many people interpret video to be "evidence" but those same people can interpret a photo to be a "claim" or perhaps some form of lower-confidence indication. Before photo manipulation was commonly known, I think photos were in a similar place as video - more trusted. Based on history, it's reasonable to expect video may follow a similar trajectory as photos to becoming less trusted in situations where it matters.

So, what happens when media types which were previously more trusted as evidence become less trusted? The same things that happened with print, audio and photos. Viewers will evaluate external cues such as the reputation of the publisher and corroborating evidence. The leading indicators that we should suspect deception will likely be similar. For example, how divergent the behavior depicted is from expectation, how contentious the surrounding context is and the existence of parties with an interest in creating such a deception.

This effect already happens with manipulation of intent through tricky video editing, for example deleting the rest of a reply to a question or even swapping in an alternate question. In the last decade I'd say the typical person is far more aware this is possible.

So, in the near-term there may be some successful deception but in the long-term I expect the potential value of creating such deceptions will diminish and we'll arrive at a new "normal" much like we have now. The biggest long-term impact may be false claims of "doctored video!" from those who were actually caught on video doing something they didn't want seen by others. But as we already see now, those pre-disposed to believe whatever is shown is false will search for indications it's doctored. Those pre-disposed to believe whatever is shown is true will search for indications it's just more confirmation of what what they already suspected. Either way, the existing reputation of the person shown, the distribution source and the pre-existing knowledge of viewers will likely be more determinitive than the media itself.