Looks like the issue is that Goodhart's Law is at play here, and there's no infallible algorithm for determining scientific merit.