The AI voices used in this test were Solaria, Natalie, Sheena, Mai, Hayden, Kevin, Jun and Asterian. They are all in Dreamtonics Synthesizer V. I did nothing to tweak them, this is exactly how they sang the midi file from BIAB. Of course, you can modify any one of them, but I wanted to see how fast this could be done.

As far as loosening up the tightness, there are several ways to do that. You can either randomize the MIDI file before sending it to Synthesizer V, or (as I prefer) alter the WAV files Synthesizer V sends back to the DAW. This can be done in Melodyne, which has an audio timing randomizer, or put a small plus or minus delay on the tracks (or both.) I did neither of these in this example, but if I had it would have sounded a bit more realistic and human.