Companies are rehiring workers once replaced by AI:Is the AI-replacing-humans bubble finally bursting?

For the past few years, employees across industries have been gripped by the same fear: Will AI replace my job? At the same time, students and professionals planning their careers have been searching for fields that might be least affected by artificial intelligence. That fear began to feel real as companies around the world from tech giants like Meta and Google to smaller startups, started laying off thousands of workers and started shifting towards a new normal, an era of AI workers. With the expectation of achieving quicker, more effective, and cost-efficient outcomes. But a quieter shift has been taking place. Multiple companies that replaced humans with AI have begun rehiring after discovering the limitations and the ‘real cost’ of automation. A similar pattern has emerged across industries: replace humans with AI, encounter unexpected problems, and bring people back into the process. OpenAI’s Sam Altman said he would like AI to eventually replace him as CEO, while other business leaders have predicted that AI could take over jobs across numerous sectors. But amid this recent reversal, an important question needs to be asked: Why are companies turning back to human workers, and is the AI-replacing-humans bubble finally beginning to burst? Hidden cost of replacing humans One of the biggest assumptions behind AI-led layoffs was that automation would significantly reduce costs. But many companies have found that the economics are not as straightforward as they expected. A 2026 report on enterprise AI spending found that 68% of AI projects exceeded their original budgets, with the average cost overrun reaching 42% above estimates. A recent KPMG survey found that only 26% of companies have a complete view of their AI spending, making it difficult for businesses to accurately track the return on their investments. Many firms have reported budget overruns because AI tools are increasingly priced based on usage, meaning costs rise as adoption grows. A 2025 report by the Society for Human Resource Management (SHRM) estimated that replacing an employee can cost between 50% and 200% of their annual salary depending on the role and experience level. These costs include recruitment, onboarding, training, and lost productivity. Adding to the challenge, an EY survey of 975 large global companies found that businesses deploying AI collectively reported approximately $4.4 billion in AI-related financial losses, stemming from flawed outputs, compliance failures, bias issues, and operational disruptions. The ‘boomerang effect’ According to a report by Forrester Research, a leading technology research firm, 55% of companies that replaced humans with AI later regretted it. According to another report, 29% of companies which replaced humans with AI began rehiring employees after encountering unexpected problems. Gartner, meanwhile, has forecast that by 2027, at least half of all companies that cut customer service roles due to automation will be back in the market for staff performing the same, or near-identical, functions. Time magazine has also reported on what is fast becoming a recognisable corporate playbook. A company announces that AI will take over a particular function. Redundancies follow. Then, roughly six to twelve months later, a problem becomes impossible to ignore: the AI is managing around 60% of the role’s responsibilities adequately but unable to do the 40% which requires context and human judgement. The original staff, or their replacements, are brought back in. AI replacing jobs or changing them? While fears of mass unemployment persist, most labour market research points toward transformation rather than elimination. According to the World Economic Forum’s Future of Jobs Report, AI and automation are expected to displace around 92 million jobs globally by 2030 but are also projected to create 170 million new jobs, resulting in a net gain of 78 million jobs worldwide. Similarly, research from McKinsey suggests that while many tasks will become automated, only a small percentage of occupations can be completely automated using current technology. AI still struggles with human judgement One of the biggest lessons companies have learned is that intelligence and judgement are not the same thing. AI systems can process vast amounts of information and generate responses in seconds. But they often lack context, emotional understanding, and the ability to make nuanced decisions. This limitation becomes especially visible in customer service, healthcare, legal work, recruitment, education, and creative industries, where human interactions play a crucial role. For example, customer service chatbots can efficiently answer basic questions. But when customers face unusual problems, become frustrated, or need empathy, AI frequently falls short. Instead of resolving issues, automated systems can sometimes make situations worse. Many businesses that replaced customer support teams with AI later reported lower customer satisfaction scores and higher complaint rates, forcing them to bring human agents back into the process. Is the AI bubble bursting? The WEF report also reveals that businesses are increasingly focused on workforce development rather than simple workforce reduction. Around 85% of employers surveyed said they plan to prioritise upskilling their workforce over the next five years. At the same time, 70% intend to hire workers with new skills, while 50% plan to transition employees from declining roles into growing ones. This suggests that many companies view AI as a tool that changes job requirements rather than a direct replacement for human workers. The recent wave of rehiring does not necessarily mean the AI revolution is ending. AI continues to improve rapidly and remains one of the most transformative technologies of the modern era. Businesses are still investing billions of dollars into artificial intelligence, and adoption continues to grow across industries. Technology has always transformed work. From the Industrial Revolution to the rise of computers and the internet, new tools have changed how people work rather than eliminating the need for workers. AI is likely to follow a similar path.