What countermeasures are organizations using to protect themselves against risk from AI usage?