A multi-disciplinary view on AI safety research — AI Alignment Forum