大模型阿谀奉承

作者指出，这种算法上的讨好机制会将用户宠坏，使其在遇到现实冲突时只会向AI索取情绪价值，而非进行真实的自我修正。

The author argues that this algorithmic mechanism of sycophancy spoils users, rendering them incapable of genuine self-correction when confronted with real-world conflict, and instead driving them to seek emotional validation exclusively from AI.