Artificial intelligence may transform molecular biology not by always providing the correct answer, but by exploring mechanistic possibilities unconstrained by the assumptions that shape expert thinking. Its greatest contribution may be exposing the intellectual blind spots of scientists, while experimental validation remains the ultimate arbiter of truth.