First the layoffs, then the rehiring: AI’s expensive round trip
- Josephine Tan
Ford spent the past three years hiring roughly 350 veteran engineers to fix quality problems its AI-driven systems had failed to catch. The push helped the automaker top JD Power’s 2026 Initial Quality Study for the first time in 16 years, though Ford remains the most-recalled carmaker in the US this year.
The organisation has been candid about what went wrong: it had assumed that feeding its design requirements into AI would be enough to produce a high-quality product, only to find that many of its most experienced engineers had left before that knowledge could be captured. The returning engineers now mentor junior employees, rebuild the data pipelines that trained Ford’s AI, and run design reviews to catch faults early.

For Vikas Aditya, CEO of technical hiring platform HackerEarth, the episode says less about the limits of AI than about how organisations value the people who work alongside it.
“I don’t think Ford is really a story about AI being weak,” he tells HRM Asia. The organisation let go of the engineers who knew what “good” looked like, he argues, and once they were gone, there was no one left to tell the AI when it was getting things wrong.
A familiar pattern
Ford is not the first name to reverse an AI-first bet. Klarna, the Swedish fintech, said in 2024 that an AI assistant was handling the workload of some 700 customer service agents. By the following year, its chief executive acknowledged that the organisation had leaned too far on automation at the expense of quality, and it began hiring human agents again to ensure customers could always reach a person. As with Ford, Klarna did not abandon AI; it settled on a model that pairs the technology with human employees.
Aditya says the underlying mistake is a common one. Many organisations struggle to recognise the expertise within their own teams, and experience is treated as overhead until something breaks. The fix, in his view, is to assess employees’ skills on a regular cadence instead of scrambling to work them out only when layoffs are on the table—particularly for newer capabilities such as AI fluency, which he says are difficult to gauge through a self-assessment or a manager’s review.
The shift is also changing what hiring processes should measure. AI already writes a substantial and growing share of production code: in Sonar’s 2025 State of Code research, developers estimated that more than 40% of the code they commit is AI-assisted, even as the overwhelming majority said they do not fully trust that output without checking it.
READ MORE: The hardest question in AI transformation: “So what?”
That is precisely the gap Aditya points to. Most technical interviews still ask candidates to write an algorithm from scratch, he notes – the one task AI already handles well. The more useful test now is whether a candidate can read code they did not write and judge whether it is actually correct, because AI output tends to look finished even when it is quietly wrong in ways only an experienced engineer catches. Drawing on assessment data across hundreds of organisations, he says the emphasis is moving toward judgment over typing, with employers increasingly screening for problem-solving and AI fluency.
That change carries a cost that HR teams are already absorbing as junior developers lean on AI, senior engineers report spending more time reviewing machine-written code—a dynamic Aditya argues makes experienced engineers more valuable, not less.
Whether the knowledge comes back
For organisations that have already cut experienced employees on the assumption that AI would cover the gap, Aditya was more guarded about the path back. Some knowledge can be recovered by rehiring, as Ford did, but he says it usually costs more than the original savings, and some expertise is simply gone once people have moved on.
His recommended starting point is a clearer read of what a workforce already has: mapping the skills an organisation relies on to identify the real gaps, then upskilling employees to work with AI. The organisations that do this consistently, he says, are the ones least likely to find themselves rehiring in a panic later.
For employers in Asia-Pacific, the tension is intensifying. The region is among the fastest to adopt AI, yet skills remain the binding constraint. In JLL’s 2026 Future of Work Survey, respondents in Asia-Pacific cited the AI skills gap as their main obstacle more often than those in any other region. Aon’s Human Capital Trends Study points the same way: while roughly three-quarters of organisations in Asia-Pacific have deployed or piloted AI, only about one in five say they can recruit and retain the AI-skilled talent they need – a gap between adoption and capability that HR teams will be left to close.


