AI layoffs bring immediate karmic retribution! A company that cut 55% of its workforce apologizes for the harm, and Ford takes three years to rehire 350 veteran engineers.

Ford vehicle hardware engineering vice president Charles Poon admits that the company’s wrong assumption was that simply feeding design requirements into AI would produce high-quality products, and over the past three years it rehired 350 senior engineers to patch the quality-control gaps. Orgvue, an organization design software company, surveyed 1,163 senior decision-makers and found that 39% had laid off staff after implementing AI, with 55% of them believing the original decision was wrong. IBM and Australia’s Commonwealth Bank of Australia have also gone down the same path.
(Backgrounder: Is the tech industry afraid to blame AI anymore? Robinhood cut jobs by 10%, and internal messages revealed Silicon Valley’s “new pretext” for downsizing)
(Additional context: Coinbase: Over 95% of code has been written by AI—by 2030, agent work will be equivalent to the output of 100k employees)

Table of contents

Toggle

• Ford took three years to bring people back
• IBM got stuck at that final 6%
• Rehiring costs more than layoffs

Key takeaways

• Ford spent three years rehiring 350 senior engineers to fix vehicle issues that its AI quality-control system couldn’t catch
• Orgvue’s survey shows 39% of companies had laid off staff because of AI, and among them 55% later thought the decision was wrong
• In the Careerminds survey, 30.9% of companies found that the money spent on rehiring exceeded what they saved from layoffs

As reported by Forbes, Ford spent three years rehiring 350 senior engineers just to fix vehicle problems that its AI quality-control system couldn’t catch, and the payback came quickly. Charles Poon, vice president of vehicle hardware engineering, told the media bluntly that the company got its judgment wrong: “We mistakenly believed that once we introduced artificial intelligence and fed the design requirements we had into it, we could produce high-quality products.”

AI isn’t useless—it’s that the company got wrong which part of the work AI could handle.

This isn’t a single company’s failure. In an international survey commissioned by organization design software company Orgvue and carried out by Vitreous World, 1,163 C-suite executives and senior decision-makers were interviewed. Of them, 39% admitted they had laid off staff due to implementing AI, and 55% of those people later said the layoff decision was wrong.

The same survey also has an even more awkward figure: 23% of companies admitted their layoff decisions were based on broad assumptions about AI capabilities, rather than itemizing what the people being laid off were actually doing day by day.

Cut people first, then study what employees do.

Ford took three years to bring people back

Of the 350 people Ford rehired, some were former employees and some had originally been working on the supplier side. The media calls them “gray beard” engineers. Poon said the core problem wasn’t the technology itself, but the training data. The company’s most experienced group of engineers had already left Ford before their knowledge was ever documented.

In other words, what AI couldn’t learn wasn’t too difficult—it was simply never written down. Those veteran engineers, who could hear the odd noises that came only from three decades of experience, and spot assembly tolerance issues, had never been put into any design-requirements documentation—so they couldn’t enter the training data.

Bringing people back worked. In its latest J.D. Power survey of new-car quality (Initial Quality Survey), Ford took first place among mainstream automakers—its first top finish in 16 years. CEO Jim Farley said that the cumulative decline in warranty and recall costs created a “very real several-hundred-million-dollar” cost tailwind.

IBM got stuck at that final 6%

IBM’s version was its built-in AskHR system that handled and eliminated about 94% of routine HR requests, sounding like an automation win. The problem was the remaining 6%—cases involving ethical judgment and exception handling—where AI couldn’t deliver answers.

So IBM announced that in 2026 it would triple hiring volume for entry-level roles in the United States, covering all business units. In remarks at a forum in New York, the head of HR, Nickle LaMoreaux, was direct: “If we don’t keep investing in entry-level headcount, what will happen in three to five years?”

IBM didn’t put people back into their original positions. The new HR hires’ work is to step in when the chatbots don’t provide enough answers, correct outputs, and communicate directly with managers; entry-level software engineers write fewer routine code lines and spend more time talking with customers. It’s shifting people from “doing that 94%” to “handling that 6%.”

The example from Commonwealth Bank of Australia is the least dignified. In July 2025, it cut 45 customer service positions, saying AI voice bots had already reduced weekly call volume by 2,000 calls. The financial-industry union wasn’t convinced. It took the case to a labor arbitration body, arguing that call volume was actually rising—meaning the bank had to ask customer service staff to work overtime and require team supervisors to personally answer the phones.

In August 2025, on August 21, the bank withdrew the layoffs, issued a public apology, and back-paid salaries. The statement was carefully worded: “An initial assessment indicated these 45 positions were no longer needed, but it did not sufficiently consider all relevant business factors. This error resulted in these positions not constituting redundancy.” In other words, the company misestimated.

Rehiring costs more than layoffs

Forbes columnist John Werner turned this cycle into a formula. When a company announces replacing a job with AI and downsizes headcount, after six to twelve months AI successfully takes on about 60% of the work, leaving the remaining 40% that it can’t handle. Then the company turns around and rehirs the original group.

A survey published on July 13 by workforce consultancy Careerminds gives a sense of scale. The report, based on interviews with 600 HR executives who oversaw layoffs in the past year, found that 91.6% regretted this round of AI restructuring, and only 8.4% thought the outcome was as expected. 35.6% of organizations had already rehired more than half of the positions that were cut, while 52.1% had gotten people back within six months.

• 30.9% said the money spent on rehiring exceeded the savings from layoffs
• 42.4% broke even back and forth—meaning they basically wasted effort
• 32.9% directly admitted losing key skills and expertise

Seventy percent of companies went around in circles and didn’t save any money. Workforce recruiter Robert Half’s data is even more direct: 32% of U.S. hiring managers had laid off a position because of AI, then later reopened the same or a similar position.

Research firm Forrester, in its Predictions 2026: The Future of Work, predicted that more than half of layoffs attributed to AI would be quietly reversed. It also pulled the long-term impact back down to earth: by 2030, the share of U.S. jobs truly replaced by automation would be about 6%, or 10.4 million roles; another 20% of roles would be augmented by AI rather than replaced. The vice president and chief analyst J. P. Gownder’s recommendation is to treat AI as a tool to amplify human labor, not a replacement.

Layoffs in the crypto space will feel familiar to everyone. Earlier, when Crypto.com CEO Kris Marszalek cut jobs by 12% last year, he said people who couldn’t adapt to AI would have to walk away. This month Coinbase said that more than 95% of its code has been written by AI, estimating that by 2030 the output of agents will be equivalent to 100k employees. But logically, the crypto industry is doing it because it had already stuffed in too many people who weren’t needed in the first place.

Frequently asked questions

Why do companies lay off staff using AI and then bring people back?

AI can reliably handle standardized work with documented procedures—about 60% of the tasks. The remaining parts that require experience-based judgment and exception handling were not written into training data and exist only in the minds of senior employees, forcing companies to rehire. Ford rehired 350 senior engineers within three years for this reason.

How many companies regret AI layoffs?

In an Orgvue survey of 1,163 senior decision-makers, 39% had laid off staff due to implementing AI, and 55% of them thought the decision was wrong. Careerminds interviewed 600 HR executives and the results were higher: 91.6% regretted AI restructuring, with only 8.4% believing the outcome was as expected.