The AI rehire

Ford has spent three years hiring back 350 veteran engineers to catch the quality problems its automated systems kept missing.

They are known internally as the "gray beards", some of them former Ford people and some brought in from suppliers, and they came back to run troubleshooting sessions and to retrain the AI tools that were not doing the job. One of Ford's vice presidents put it plainly, that the mistake was assuming that introducing artificial intelligence would on its own produce a high-quality product. Ford has since come top of the mainstream brands in the JD Power initial quality survey, and its chief executive has tied the fall in warranty and recall costs to hundreds of millions of dollars.

It is not an isolated case. In a survey of more than a thousand business leaders, four in ten had made people redundant because of AI, and 55% of those now say they got that decision wrong. Klarna is the other example everyone remembers, having claimed its chatbot did the work of 700 support staff before concluding it could handle volume but not nuance, and it is recruiting people again.

I do not read any of this as evidence that AI cannot do the work. I read it as evidence that very few organisations measured what their people were actually doing before they removed them, so the value only became visible once it had gone.