OpenAI’s takedown of false-front ChatGPT accounts, its EU text watermark and a court order on White House hard passes may show a wider contest over credentials

OpenAI said Thursday it shut down Russia- and Iran-linked ChatGPT clusters, one using seven fake Western journalist personas. It also plans an invisible watermark for ChatGPT text in the EU, cautioning that a missing mark “does not prove human authorship.” In late September, a federal judge temporarily restored White House hard passes for CNN, MS NOW and Politico, finding the ban likely violated due process; the administration had cited false information and national security. Read together, they may show a wider contest over credentials.

Theory - analysis