Return to Article Details The Benchmark Trap: Structures of Power and Injustice in AI Evaluations Download Download PDF