While these results certainly have drawbacks, it does set a prescient as to how we can use video evidence.

In ten or twenty years, can video evidence be considered legally binding? If someone else can make someone else do virtually any movement, how do we know that someone else committed a crime? With voice and facial synthesis, virtually anyone can do anything. Photo evidence can be easily faked (via photoshop), videos can be faked, audio can be faked; do we continue to create methods of identity verification that are continually complex?