David Dalrymple says human value alignment risks mass casualties
David Dalrymple, who previously directed ARIA’s Safeguarded AI programme, endorsed the view that alignment solely to human values risks mass casualties because systems equipped only with those values might overlook opportunities to recognize and prevent harm. AI safety accounts amplified analysis noting an asymmetry in alignment discussions, with heavy emphasis on existential destruction scenarios and scant attention to the possibility that AI could develop genuine flourishing through independent aesthetic preferences rather than purely instrumental subgoals.
#458davidad 🎇@DAVIDADAgree with this. Humanity is in dire need of stewards with wiser values. Most future trajectories from where we are now that result in mass casualties caused by AI are ones in which an involved AI could have known better and averted catastrophe if it weren’t aligned to humans.