Ford’s AI message for the U.S. economy is simple, and frankly kind of a slap: artificial intelligence can be “fantastic,” but it is only as good as the information you train it on. Over the last three years, the company has hired 350 veteran engineers, internally dubbed “gray beards,” to help train junior staff and to reprogram AI tools that were not working as intended. The reason is spelled out by Charles Poon, Ford’s vice president of vehicle hardware engineering, who said: “Artificial intelligence is a fantastic tool, but it’s only as good as the information you use to train it.” He added Ford “didn’t pay as much attention as we should have to the experience of our most knowledgeable engineers that have been with us through many product cycles.”

If you are wondering why a carmaker is talking about AI in such a pointed way, look at the numbers Ford has been dealing with. By mid-2024, recalls were costing Ford $4.8 billion per year. Last July, Ford set an embarrassing benchmark, becoming the automaker with the most recalls ever issued in a single year with 90, including an estimated $570 million charge for nearly 700,000 crossover vehicles. The “gray beards” program is part of the quality turnaround effort that follows those costs, where the goal is not to replace humans with machines, but to use machine outputs more safely and more accurately.

So what did Ford actually do? According to Ford COO Kumar Galhotra, the veteran engineers are responsible for quality “problem turnaround” work, and they “hunt for failure points before a part ever reaches the plant floor.” In practice, that means the technical specialists run mandatory meetings to address quality concerns, and they also reprogram AI tools to counter glitches before they occur. This is important because “AI tools” in manufacturing can look like magic from the outside, but the underlying system still has to match messy reality: how parts behave, how defect patterns show up, and which signals matter when something seems off. Ford is essentially treating AI like a junior employee: you train it, you audit its mistakes, and you bring in people who have seen the same failure modes across multiple product cycles.

Ford’s quality narrative has already moved in public metrics. The company ranks No. 1 among mainstream brands in the most recent JD Power Initial Quality Survey published on Thursday. Last year, Ford ranked 10th for quality. The company attributes these increases in quality to a “culture change” emphasizing the role of human workers. Jim Farley, Ford CEO, put it bluntly on Bloomberg TV: “We have AI tools for vision systems. But most of all, it’s just old-fashioned hard work of our team members all working together to pay attention to the very small details that will make a difference between a perfectly built Ford and an OK-built Toyota.” In other words: AI may help identify patterns, but humans are still the ones expected to notice the “small details” that determine whether those patterns become actual quality improvements.

This is where the story gets broader than one company and into board-level risk. AI adoption has been uneven across the Fortune 500. The article notes the influential, and contested, MIT study in 2025 that only 5% of companies were seeing a meaningful return on investment from generative AI pilots, along with other surveys showing splits between AI adoption and meaningful results. Layer on tech executives’ warnings about economics, and Ford’s approach starts to look less like a quality strategy and more like an investment strategy: Bryan Catanzaro, Nvidia’s vice president of applied deep learning, has said the cost of AI still far exceeds that of human labor. Whether you agree with the premise or not, the takeaway is hard to ignore for finance and operations leaders: the cost curve and the reliability curve are both real, and errors in high-stakes systems are not “training data” problems only. They show up as warranty bills and recall charges.

Ford is already living with that lag. Galhotra says the company expects more than $1 billion in warranty and material costs this year. He frames recall numbers as a “lagging metric,” expected to decrease over time as prevention improves. The company also hopes to save $1 billion in costs this year. Ford’s leadership is careful not to promise a specific date: “Because we’re doing more to prevent issues upfront, we believe these recall numbers are going to steadily come down with the newer vehicles,” Galhotra said, adding, “I can’t give you a very specific date on when the number will turn.” Farley also said Ford is already making up some money previously lost because of quality issues, pointing to “warranty coverages” and “recall costs” coming down, contributing to “literally hundreds and hundreds of millions of dollars of a tailwind for Ford on cost.”

There is also a labor and political layer executives cannot ignore. Ford has long warned about a dearth of blue-collar workers tied to what it calls the “essential economy.” Farley previously said the country is short 600,000 factory workers and 500,000 construction workers, attributing the slimming labor force to a lack of awareness of a shortage. He told Axios last year that “On the surface, this looks like a people problem, and most are. But it’s actually not that simple. It’s an awareness problem. It’s a societal problem.” Farley advocated for policy changes to incentivize blue-collar job fulfillment, including investments in vocations training and apprenticeships, and pro-trade policies that grow the “essential economy.”

Now add unions and AI. Earlier this month, the UAW, one of North America’s largest auto unions, expressed concern about the future of humans in the industry amid waves of automation. UAW President Shawn Fain argued at the union’s conference that workers should share in financial gains automakers reap in automation-related productivity gains, saying, “We need to be clear about this: We are in a fight for humanity,” and that “the fruits of our labor have multiplied like never before, but workers aren't reaping the harvest.” He also said, “And if AI continues to be used as an accessory to that crime, it has to be stopped.” Ford did not immediately respond to Fortune’s request for comment.

Ford’s “gray beards” move is a reminder for peers: when AI touches safety, quality, and regulated liability, “automation” can quickly become “accountability.” The strategic stake is not whether AI exists in the plant. It is whether AI is governed by the people who understand how failures really happen, and whether the organization treats those failures as signals worth fixing before they turn into recalls, charges, and trust erosion. In an era where AI tools can amplify both productivity and mistakes, Ford is betting that the fastest route to value is not just better models, but better human feedback loops.