Disrupting malicious uses of AI by state-affiliated threat actors

Spoofing and false flag attacks are the name of the game here. We don't actually know if the election bots in 2016 were Russian, just that American agencies selected Russia as the casting target for the big public accusation. Authoritarian regimes regularly blame Western intelligence agencies for all sorts of domestic problems in order to legitimize their regime and deflect blame for what is actually an embarassing internal conflict, it wouldn't be surprising to see that it often goes both ways.

Notably, Microsoft contributed substantially, even though Microsoft itself is a state affiliated threat actor. Microsoft could have been all 5 of these and I doubt OpenAI would have had any chance of finding out themselves.