Blake Lemoine violated “employment and data security policies,” Google spokesperson Brian Gabriel told The Washington Post. In June, Lemoine told the publication that he believed LaMDA, the company’s chatbot generator, had achieved sentience. (Washington Post)

Talking point: Google said there’s no evidence that the machine contains a ghost, and plenty to suggest it hasn’t come to life. It’s just pattern-matching, AI experts say—a sophisticated auto-complete, but not one with feelings. Lemoine talked philosophy with LaMDA, and claimed its grasp of the material was more sophisticated than graduate-level conversations he’d had with Ivy League scholars. While he’s busy worrying about the high-mindedness of that code packet, a more physical meeting of human and machine went awry last week in a chess match between a robot and a seven-year-old boy. “The robot broke the child’s finger,” said Sergey Lazarev, president of the Moscow Chess Federation. “This is of course bad.” But does the robot feel bad?