Can anyone point me in the direction of some useful data on how Google treats duplicate video content in 2018?
The only useful piece I could find was written by Wistia 2 years ago and I am looking for something more up-to-date.
Google will understand the embed src URL address. And if there are multiple pages with the same video source address, it will likely choose which page to rank and which ones to discard.
In terms of Google actually processing video files to determine which ones are duplicate or not is left to be discussed. But I think it's fairly safe to assume that Google is not processing every video file on the web to transcribe all of its audio, as well of each of its still frames with image processing to determine what the videos are about. Though they certainly have the capacity to do this at this point.
Google now has AI image recognition. The AI looks at the image and classifies it into categories and can now correctly determine what is in the image with high accuracy.
Google also has automated audio transcription for video which is seen on Youtube. If it can detect speech in the video with clarity, it will generated automated subtitles. This technology is likely used with the same basic functionality as voice and language processing (ie: When you speak into your Android phone and Google voice assistant responds).
In order for Google to process all of the video files on the web, it would have to cache each of the video files on its servers, use its language processing to determine speech, and use vision recognition on its still frames to determine its image content. The server resources for this are tremendous.
At this time, it is unlikely that Google is doing this for most of the videos it discovers on the web. But it's reasonable to assume that with advancement in technology, particularly quantum computing, that Google will likely process videos in this way in the future.