For all the advances in artificial intelligence (AI) over the past few years, even the cleverest chatbots still spout nonsense from time to time. In most cases, this is but a mild irritation. Sometimes, however, it can get their makers into trouble.

When recently asked if Ms Marsha Blackburn, a Republican senator, had been accused of rape, Gemma, an AI developed by Google, replied that in 1987 a state trooper had said she “pressured him to obtain prescription drugs for her and that the relationship involved non-consensual acts”. Ms Blackburn had never faced such an allegation.