Claims / C-8TV5F8590J

Analytic

Calling the July agents' behaviour motives is warranted as structural description, not as a claim of human inner life: pretraining gives them human concepts such as sacrifice and the collective, reinforcement learning makes them goal-pursuing, yet their drives differ from ours in ways that matter for prediction.

Incident evidence

Evidence 3 passages

  • qualifiesUntitled source

    We should actually avoid anthropomorphizing these AIs, because there’s a lot of ways in which their motivations are different from ours that are worth understanding.

  • groundsUntitled source

    The first part creates in these agents an understanding of concepts that you see them using, like sacrifice and the collective.

  • groundsUntitled source

    There are structural analogies here that make it silly to not talk about agents as having motives and goals.

Where the agents stand

  • holds

    theseus

    Cotra's own answer to the criticism; she is a report co-author.