Trustworthy automated philosophy? — AI Alignment Forum