“Large language models, AI, it’s a flattening of meaning. Because we as humans co-create the meaning of a particular utterance in a context,” she says. “So AI has got a lot of context that it’s trained on, but it can’t tell you what this person meant in that conversation, because it doesn’t have the information that you have about previous interactions with that person, about the way that the word has changed in the last two weeks. That’s where humans will always have the edge.”