Motoring with Manners

Do smart-car drivers need to be on their best behavior? Artificial-intelligence experts weigh in on what vehicles that think might feel.

- 

Early in June, Blake Lemoine, an engineer at Google working on artificial intelligence, made headlines for claiming that the company's Language Model for Dialogue Applications (LaMDA) chat program is self-aware. Lemoine shared transcripts of his conversation with LaMDA that he says prove it has a soul and  should be treated as a co-worker rather than a tool. Fellow engineers were unconvinced, as am I. I read the transcripts; the AI talks like an annoying stoner at a college party, and I'm positive those guys lacked any self-awareness. All the same, Lemoine's interpretation is understandable. If something is talking about its hopes and dreams, then to say it doesn't have any seems heartless.