I think the article really makes the point that it doesn't matter. The videos can LITERALLY have a watermark on them from the AI software, and people just dgaf. The battle is lost before it begins.
As with things like quotes and claims that could always be fake, it's back to the old journalistic practices of verifying sources with a second source. But that means being ignorant of things that were not being investigated by journalists, which creates a different filter bubble.