Visuals that are generated by artificial intelligence $($AI$)$ and$/$or deepfakes can be challenging to spot with specific online tools. It's becoming more difficult to detect a deepfake, but they do often share some of the following characteristics:

glitches, distortions and other misformed elements, especially with hands, teeth, skin and face movement

language that isn't clear in a video $($muffled$)$ or mouths that seem to not form words correctly

backgrounds that are surreal or hyper$-$bright in color or texture

small changes in lighting from the background to the foreground that make it look like it has been Photoshopped