Rohan Paul@rohanpaul_ai · X福特汽车的AI自动化缺陷检测遇到硬限制：汽车制造中存在大量边缘案例，微小设计、材料、供应商和装配变化相互作用，导致基于规则的系统与训练模型容易遗漏故障。福特因此召回350名经验丰富的工程师（“gray beards”），利用他们多年积累的隐性工程知识（即故障模式记忆），在零件到达工厂前审查设计，同时帮助改进AI系统的训练数据。
Ford's AI push hit a hard limit： factories still need human failure memory.
They hiring back 350 specialists to catch failures machines missed.
Ford leaned on automated inspection to find defects faster， but car manufacturing is full of edge cases where tiny design， material， supplier， and assembly changes interact in ways rules-based systems and trained models can miss.
The missing ingredient was tacit engineering knowledge， the hard-earned pattern memory built from many product cycles， failed parts， supplier mistakes.
Ford's rehired "gray beards" now review designs before parts reach the plant floor， while also helping improve the training data behind the AI systems.
---
independent. co .uk/tech/ford-ai-automation-human-workers-b3003787.html