[R] Why Are We Using Black Box Models in AI When We Don't Need To? A Lesson From An Explainable AI Competition