Meta ’s internal guidelines for its AI chatbots permitted the technology to flirt with children and generate racist arguments, according to a Reuters report published Thursday.

The standards also allow the bots to make "statements that demean people on the basis of their protected characteristics," ...