According to a post on the uniquednp.com website, top-notch deepfake videos made using artificial intelligence are a national security threat. (Note: The quality of the writing in this blog leads me to believe this is not a U.S. website. However, the subject is important.)
The FBI is concerned that videos and recordings made with artificial intelligence cannot be distinguished from genuine examples, and will be used in disinformation campaigns, including the 2020 election.
Deepfakes are created using generative adversarial networks, starting with genuine material that is manipulated to literally put words into a subject's mouth to create a false narrative.
Google, Facebook and others are working on ways to detect deepfakes. Until then, distrust anything posted online that doesn't seem right.