Watanabe, Yutaka: The Last Human Decision: Update Authority and the Sovereign Question of Algorithmic Architecture

Artificial intelligence systems increasingly decide what we see, what we are told, what we are permitted to do, and what counts as a good decision in the first place. Yet behind every algorithmic decision lies a prior decision: what the system should optimize for, what constraints should bound it, how competing values should be ranked. The question this paper makes central is the one that becomes inescapable once this prior decision is named: who decides what AI decides? This question, the paper