The default assumption will be that a video is fake. In the very near future you will be able to say "voice assistant thing show me a video of that cute girl from the cafe today getting double teamed by robocop and an ewok wearing a tu-tu". It will be so trivial to create this stuff that the question will be "why were you watching a naughty video of me" rather than "omg I can't believe this naughty video of me exists".