[x62Bh7948f]

I don’t think Google’s AI is close to being sentient. Being able to generate “convincing” phrases from prompts is not that relevant to judge whether it is “conscious” or not. As a simple elimination game I would propose a series of word games: explain basic rules and see if it gets it right. Explain mistakes and see if it can learn on the go. A sentient AI with the intelectual level to claim “being afraid of being turned off” can surely learn a couple rules for a kid’s game. If after that it keeps going, let’s try having it explain the game to someone else. Then I would think of more elaborate rules, asking for motivations and trying to figure out if there are any changes in its self perception over time. It was very unfortunate and unprofessional of Lemoine to make such claims without putting proper reasoning to it.