The Department for Science, Innovation and Technology has published binding regulatory guidelines for developers in London and Cambridge, requiring mandatory safety testing for large-scale artificial intelligence models.