This Reddit thread shows the concept in a concise way:

https://www.reddit.com/r/ChatGPT/comments/1c77pr8/microsoft_image_to_video_is...


 Basically we can now take any front facing photo of a human and turn that into a rather realistic video of that human speaking.  This is going to run rampant on video-chat for the elderly.  Scammers are going to love this software as they make real-time videos of grandkids asking grandpaerents for emergency cash.