These things are deliberately constructed to mimic human language patterns, if you're trying to determine whether there is underlying sentience to it, you need

These things are deliberately constructed to mimic human language patterns, if you're trying to determine whether there is underlying sentience to it, you need to be extra skeptical and careful about analyzing it and not rely on your first impressions of it's output. Anything less would be a level of stupidity not fit for a Google AI researcher, which considering that he was fired is apropos. That he keeps going on about it after his 15 minutes are up is not proof of anything except possibly that besides being stupid he also stubborn.