ABSTRACT
How intelligence is defined has a cascading effect on how artificial intelligence (AI) is defined, developed, managed, and governed. Intelligence is an ambiguous concept—to identify something as being intelligent as opposed to unintelligent is a judgement based on varying norms and values. I propose a theory of intelligence that embraces that variability by conceptualising intelligence as value-dependent cognitive performance. I develop my theory into an analytical framework which I use to queer two influential theories of intelligence: John Carroll's three-stratum model of individual human intelligence and Alan Turing's ontology of AI. Across both of those theories, intelligence is defined by many implicit cognitive, normative, and performative qualities that I critically call into question. My exploratory questions reveal an expansive agenda for future ontological, critical, and practical studies of intelligence and AI.
