I remember I once saw a post here about research at (I think) MIT where they were able to use shadows in photographs to “see” what was behind the camera and “see” around corners. It was really rough and didn’t really look like very much but I think about that research all the time. What happens when that technology improves and every photo that has been posted online is suddenly a photo of way more than the photographer intended to share?

You'll probably be interested in https://www.bellingcat.com/news/2020/04/15/did-these-syrian-...

You’re probably thinking of the Dual Photography work from Stanford (and all the follow-on work): https://graphics.stanford.edu/papers/dual_photography/