Research and classroom pilots underscore that teacher‑provided grading continues to produce learning gains that AI alone does not replicate, even as instructors experiment with AI to scale feedback. Educators and ethicists argue that human graders supply contextualized judgment, relational coaching and developmental feedback critical for student growth; at the same time, pilots show AI can accelerate revision cycles by generating isomorphic practice problems and annotated solutions. Institutions are testing hybrid workflows that require teacher oversight while using AI to surface patterns, personalize practice, and free instructor time for higher‑order mentoring.