With AI having its moment, I’ve been thinking about the value of individuality. Chatbots are now clearly exhibiting personalities of a sort—they are, for example, capable of quipping (or at least, generating something that we might perceive as a quip).
One of the curious things that might happen as the AI becomes more personable is that it becomes in some ways and not in others. How does its personality hold together, relative to its knowledge? One of the most interesting aspects of this is that the more personable an intelligence becomes, so too the less omniscient—the bots can’t yet deliver all possible responses to a prompt at once (because there are infinite possible responses). To deliver something in language is to choose one option from many possibly legitimate (or illegitimate!) choices.