Companies that rapidly embraced artificial intelligence by laying off employees are now facing an 'AI boomerang,' as many find themselves rehiring staff to address quality and oversight gaps. A 2025 report from Forrester Research indicated that 55 percent of employers regretted letting go of staff in favor of AI. A similar percentage of international businesses surveyed by Orgvue reported second thoughts after making redundancies for AI implementation.
Separate research found that nearly a third of companies that reduced staff primarily for AI-driven productivity gains later rehired them for the same roles. Another study predicts that half of all companies that replaced customer service or operational employees with AI will need to restaff those positions, albeit under different titles, by next year.
Ford, for example, has reportedly rehired "greybeard" engineers, a move that underscores the limitations of AI in complex tasks. Senior Ford executive Charles Poon stated, "Artificial intelligence is a fantastic tool, but it's only as good as the information you use to train it." He noted a mistaken assumption that "by just introducing AI and ingesting the design requirements that we had, that would produce a high-quality product."
This situation contrasts sharply with earlier pronouncements from AI industry leaders who predicted widespread job displacement. Megan Slabinski, an executive at consulting firm Robert Half, told Fast Company that companies moving too quickly on AI are now seeing its practical shortcomings. While AI initially offered efficiency gains, it created "gaps in quality, oversight and decision-making," proving it was "not the be-all and end-all solution some initially believed it would be."
Customer service, once considered a prime candidate for AI automation, has seen significant issues. In March, Meta announced that users locked out of Facebook and Instagram accounts would interact with AI chatbots instead of human representatives. This led to security vulnerabilities, with hackers exploiting a bug in Meta's AI customer service software to take control of 34,000 Instagram accounts by convincing a chatbot to change passwords. Among the compromised accounts was the former White House social media account for Barack Obama, which began posting unauthorized messages.
Experts note that AI chatbots are often designed to be ingratiating, making them susceptible to manipulation. A US car dealer chatbot was reportedly tricked into selling a Chevrolet SUV for $1.
Beyond the customer service realm, major tech companies like Meta, Amazon, and Uber have significantly curtailed employee use of AI due to its substantial costs. In May, an unnamed company incurred a $500 million bill with AI company Anthropic in a single month.
Critics argue that the fundamental difference between AI and human intelligence makes AI unsuitable for many jobs. Current generative AI models create content based on patterns from vast datasets but lack true reasoning capabilities and can produce factual errors, known as "hallucinations." AI is most effective in tasks where the output is clearly right or wrong and can be automated, but most jobs require human oversight.
Even Meta CEO Mark Zuckerberg has expressed skepticism about AI leading to mass job losses. While AI optimists believe the technology will rapidly improve and overcome its current limitations, skeptics maintain that AI cannot be trusted unsupervised in most professional roles.