Every year, over 300,000 F-150 pickups operate their lines at Ford's Dearborn Truck Factory. This vehicle – the bestselling American car – features a complex lineup of trim levels, electrical hardware, wire harnesses, badges, fenders and seat variations.
The machine assembles the components with gloved hands, so certain new members of the company's quality assurance team are not distracted: AI-powered cameras.
Ford has deployed two in-house artificial intelligence systems called Aitriz and Maivs to detect factory defects in real time. Implemented in December 2024 and named after Spain-based creator Beatriz Garcia Collado, Aitriz uses machine learning and video streaming to catch inconsistencies in millimeter scales. Maivs, which debuted in January 2024, relies on still images of smartphones mounted on 3D printed stands to ensure that all the correct parts are properly installed on the vehicle.
Together, they help workers identify problems on the spot – before the issues turn into costly guarantee claims, recalls or rework. This is the area that Ford hopes to improve after leading the automotive industry in recalls over the past five years.
“Imagine you're an operator and you get a vehicle every minute. There are different tasks to complete, and every vehicle is different,” Jeff Tornabene, who manages vision applications at Ford's manufacturing technology development center, told Business Insider. “If the control system knows the recipe and the vision system can check it, there are useful suspenders that will make sure you have the right parts on the right vehicle.”
Cameras with AI capabilities help factory workers avoid expensive manufacturing errors. Nick Antiya for the bi
AI could be a Ford fix for costly recall issues
These tools can relieve Ford's billion-dollar headaches. This year, automakers face a record 94 recalls, affecting cars built primarily before 2023. This is the safest breaking news for major car brands posted throughout the calendar year, early August. Chrysler has the second highest recall among major automakers, with 21.
These bulletins cost millions of dollars. For example, the fuel leak recall for 694,271 units of the popular BroncoSport and Escape SUV model costs $570 million, according to Detroit Free Press.
The company's AI-supported systems have already made a huge difference to the automaker's manufacturing floors, says Patrick Frye, engineering manager at Dearborn Plant.
“It definitely helped from an operational perspective,” Fry told Business Insider. “Instead of waiting for the vehicle to finish or hit a specific checkpoint, we can improve the station's process control.”
Such speed is important. Before systems like Aitriz, Frye said many electrical issues don't surface until the final inspection. Often, it was necessary to tear the carpet or remove the seat. Also, as Ford vehicles acquire more screens, sensors and autonomous driving hardware, even slightly looser connections can cause major headaches. For human workers, the small difference between loose parts and good connections is impossible to catch.
“When you protect our hands, or protecting the noise of our general factory, you don't always hear clicks, or you can't always feel the snaps of the two connectors together,” Frye said.
Brandon Tolsma, MTDC vision engineer at Ford, said real-time AI results are often the difference between fast fixes and carpet ripping issues.
Nick Antiya for the bi
“As the vehicle passes through the assembly line, it becomes more and more difficult to access some of these components,” Tolsma said. “We can't stress enough how important real-time results are to save time.”
The system is currently located in dozens of stations in North America – Aitriz is 35 stations and Maivs is close to 700. Maivs is limited to still images, but Aitriz's live video feeds provide more accuracy and adaptability, especially in the case of occlusion (such as when workers walk through camera frames or block electrical connections).
Previously, due to basic electrical contact, some failed connectors passed the final inspection, Torsma said. However, if the plugs were not fully seated, they can later be loosened in real-world driving. Ford believes that AI cameras are catching the millimeter-off subtlety that factory workers may miss, even if they have extensive experience, hidden behind sheet metal and carpet.
“It can be a bit hard to see, especially if it's a little far from the electrical connector, as the operator is about to work right away,” Tolsma said. “But with vision, a powerful camera can see the difference quite clearly.”
Starting a smarter assembly line
Analysts are cautiously optimistic. David Wyston, Morningstar analyst covering Ford, told BI that AI has shown long-term promises, but it is still in its early stages.
“I think if AI is done effectively, we can reduce recalls over the next decade or more,” he said. “But for now, there is no guarantee that it will work. In theory, we need to reduce the costs of warranty and recall, not just in manufacturing, but also in design and engineering.”
For now, Tornaben says the vision tool is intended to support manufacturers.
As models become more complex, AI systems like Aitriz and Maivs could potentially support plant operators implementing more sophisticated sensors, computer chips and processors to aid Ford's autonomous driving ambitions, Tornaben told BI. It also helps to get through the noise of the factory so that each vehicle is built as intended.
“This is not an endpoint for us,” Trnaben said. “This is just the beginning of a plan that requires this technology to be a quality game changer.”
