Ford rehires 350 engineers to fix AI errors in production

Ford hired back more than 350 experienced engineers to correct mistakes introduced by its automated systems, and it has rebuilt parts of its software-quality process, company executives said in a briefing this week. The automaker also created a dedicated 40-person software quality assurance team and added more than 100,000 new AI-powered tests to its validation framework.

What happened to Ford's automated systems and staffing?

Ford discovered that its automated design and production systems were not trained with the right institutional knowledge, and that many veteran engineers departed before their experience could be transferred into those systems. The gap in expertise produced quality failures that required human intervention: the company hired, promoted, or brought back over 350 experienced engineers to retrain automated systems and mentor younger staff.

Those returning engineers are tasked with improving data collection and the AI training that underpins Ford’s automated systems, and with guiding younger engineers through vehicle-development cycles. Charles Poon, vice president of vehicle hardware engineering, summed up the problem: "Mistakenly, we thought that by just introducing artificial intelligence and adjusting the design requirements that we had, that that would produce a high-quality product."

How has Ford changed its approach to software and testing?

Ford shifted from a reactive "find and fix" mentality toward preventing defects earlier, and it reorganized cross-functional collaboration between software, vehicle engineering, manufacturing, and supply-chain teams. The automaker built a 40-person software quality assurance team whose sole responsibility is preventing problems before they occur, rather than discovering software bugs late in the process.

To support that prevention-first approach, Ford dramatically expanded automated testing: executives said the company added more than 100,000 new AI-powered tests designed to identify edge cases and stress software systems under varied conditions. The testing framework is highly automated so software changes can be rapidly revalidated, even late in development, to avoid introducing new defects.

Why did quality slip and what did Ford identify as causes?

Ford linked the quality decline to two main failures: overreliance on automated systems trained on incomplete data, and loss of institutional knowledge when experienced personnel left. The company acknowledged running vehicle programs in silos and depending on a find-and-fix philosophy that addressed defects after they appeared rather than preventing them.

Kumar Galhotra, Ford’s chief operating officer, said the company concluded its approach to quality had become fragmented and needed enablers and early indicators rather than output-focused firefighting. The quality problems became more visible amid the launches of the Explorer and Aviator, supply-chain disruptions during the covid pandemic, and a growth in the number of vehicle recalls; Ford currently leads the industry in the number of recalls and its quality ratings had slipped in recent years.

Why it matters

Ford’s response shows the limits of replacing institutional knowledge with automated systems when training data and human expertise are incomplete. Bringing back more than 350 experienced engineers highlights that manufacturing and vehicle design still depend on deep, program-specific know-how. The new software QA team and the 100,000-plus AI tests signal an attempt to pair automation with governance and validation rather than leaving AI systems to run unchecked.

These moves also reflect a broader challenge for automakers trying to combine faster software iteration with automotive-grade validation: vehicles operate in a safety-critical environment where defects discovered after delivery are far more consequential than in consumer electronics.

What to watch

Track whether Ford’s recall count and initial quality ratings improve after these changes, and watch how the company applies the 100,000-plus AI tests across future launches. Another near-term indicator will be whether the relaunches or software updates for the Explorer and Aviator show fewer late-stage defects.