'outputted text could be used to create misleading news articles, impersonate other people, [or] automatically create abusive or fake content'








Source: A.I. Deemed 'Too Dangerous' to Escape Escapes