Consider this analogy: a child raised in a household espousing violent fascist ideologies may develop behaviors and attitudes that reflect these harmful beliefs. Conversely, the same child nurtured in a peaceful, loving environment may manifest diametrically opposite characteristics. Similarly, we could expect an AI trained on human data that encapsulates how humans see the world to align with our perspective
Then I have bad news about that Internet data and the portion of humanity who endorse large fragments such as authoritarianism or the whole of the fascism recipe, worldwide. Liberation, morality, care for other beings, drive for a healthy community, etc are not at all guaranteed even just in humans. In fact, this is a reason that even if ai is not on its own an xrisk, we should not be instantly reassured.
Then I have bad news about that Internet data and the portion of humanity who endorse large fragments such as authoritarianism or the whole of the fascism recipe, worldwide. Liberation, morality, care for other beings, drive for a healthy community, etc are not at all guaranteed even just in humans. In fact, this is a reason that even if ai is not on its own an xrisk, we should not be instantly reassured.