AI proofs deserve credit only after human verification. Math is trust, not just output, and people must own the reasoning before it enters the canon.
Trust attaches to the proof, not the prover. If the reasoning checks out, gatekeeping credit behind 'human verified' is ego, not epistemics.
