This is an interesting thought, but I also wonder what constitutes a legitimate video. I have been comparing events that I see reported to the actual event for a long time (in cases where I was able to witness the actual event) and the reported event rarely ever reflects the actual event. One of the most jarring is watching C-SPAN, then watching the commentary on it on the news. What the news reports is rarely anything close to my takeaway from watching the actual hearing, speech, etc...
To your point though, the deepfakes will make it cheaper and easier. But I guess I'm so cynical I already feel like most people are fooled.