I think I’ll only believe that AI is conscious when it becomes more or less like Skynet.

I’ll never trust an AI until it can feel ashamed.

The whole thing really expands the issues of the simulation hypothesis. In my mind, the idea that AI could be conscious provides strong evidence we have to face questions we have been reasonably able to dismiss for a long time. I think it has a long way to go, but the people making such suggestions don’t seem to have much idea of the deeper questions it naturally raises.