I've always been curious about this - to what extent can this stuff be done in real time these days? I mean once you've successfully constructed a pop star in the lab like this, how do you handle live performances? Is lip-syncing the only option, or if you've got someone who's at least somewhat close can you do enough fixing live to keep them sounding good?
There is quite a lot you can do. Lip-syncing is still quite common but not the only option. Most of the softwares we use in the studio also have a real-time option. This real-time option is not good enough quality to use in a commercial audio release but is fine in a live performance where quality expectations are lower and there's large amounts of background noise to help cover-up the deficiencies.
https://www.youtube.com/watch?v=h3ilIfOI8fc