The problematic effects of amplified AI inference over time are straightforward to assume (although I used to be crafting this short article, a Google researcher went viral claiming he’d been placed on administrative depart right after notifying the corporate that one of its AI chatbots—