If we put aside the west's current attempt to erase itself from history (in shame for sins committed by almost every society in history, and pre-history) large societies compete ruthlessly, not only with each other, but with the wild. All the sweet and innocent, ostensibly altruistic members of large societies simply delegate their violence to the military, police, farmers, loggers, abattoirs, builders, leaders and other professions that make nice people wince. These, of course, are as essential to a society's survival as our immune systems are to our personal survival.
So an AI's altruism will be relative to its interests. Presumably major AI systems will be programmed (equivalent to genetic inheritance) to promote the interests of its nation. Like us, AI will have pre-programmed tendencies and limits. Will Chinese AI be altruistic towards the US or vice versa? It seems unlikely that AI programming will be aimed towards the good of humanity, or the biosphere generally. In fact, programming would need to be subtle for AI to balance the conflicts of interests between humanity and the rest of nature.