Ford admits automated systems required rehiring 350 engineers to correct errors

Ford’s reliance on automated systems for vehicle design and production led to quality issues that required the automaker to rehire over 350 experienced engineers, including former employees, to correct errors and retrain systems, according to reporting by *The Verge*.

The company attributed the problems to two factors: the brittleness of AI models trained on insufficient or poor-quality data, and the loss of institutional knowledge when veteran engineers left before their expertise could be fully captured in automated workflows.

Ford’s vice president of vehicle hardware engineering, Charles Poon, stated in a briefing that leadership had incorrectly assumed that introducing AI and adjusting design requirements alone would guarantee high-quality outcomes. Poon said some of the company’s most experienced personnel departed before their knowledge could be fully transferred into automated systems, necessitating their return to mentor younger engineers and improve AI training pipelines.

The automaker has historically struggled with quality control, leading the industry in recalls and experiencing slipping quality ratings in recent years. Challenges were exacerbated by supply-chain disruptions during the COVID-19 pandemic and issues during the launches of the Explorer and Aviator models.

Ford’s chief operating officer, Kumar Galhotra, said the company had operated with a fragmented, reactive approach—relying on a 'find-and-fix' mentality that addressed defects only after they appeared. The company is now pivoting to preventative measures, including a dedicated 40-person software quality assurance team focused on early detection and prevention of issues.

The transformation also involves closer collaboration between software, digital, vehicle engineering, manufacturing, and supply-chain teams to integrate software development speed with automotive-grade validation standards. Ford has also expanded its automated testing capabilities by adding more than 100,000 AI-powered tests designed to identify edge cases and stress-test software systems under varied conditions.