Google DeepMind is worried about what happens when millions of agents start to interact

MIT Technology Review 发布的媒体报道：Google DeepMind is funding research into the potential dangers of situations where millions of different AI agents interact with each other online. According to Rohin Shah, who directs the company’s AGI safety and alignment research, the mass-market arrival of agents that can carry out tasks without human oversight and follow instructions given to them by other…