Is there something useful we can ask governments to do for AI alignment?

From Stampy's Wiki