I'm concerned that there's no group of actors who value safety over self-interest with enough power to implement a ban, and the key actors have immediately show

I'm concerned that there's no group of actors who value safety over self-interest with enough power to implement a ban, and the key actors have immediately shown themselves to be compromised. It seems the key actors with control of the technology and resources to develop the technology are being absorbed by capitalistic institutions, with interests in safety compromised by profit motive. In a game-theory sense, to voluntarily implement the ban would be suicide. States are concerned that other states will gain access to AGI with decision making power (most likely) gain a competitive advantage, with an outside chance of a nuclear moment - an AGI singularity under state control. Other states cannot be relied on to cooperate in the current global environment, there is risk to not remaining competitive in AI. I'm not seeing the dynamics which allow this ban to actually occur. How could it be made to happen? Who needs to meet, and what needs to be agreed upon?