AI is replacing entry-level jobs…or is it?

June 24, 2026

Many organisations are reducing hiring for entry-level positions, citing AI’s ability to automate tasks that once required junior employees. Administrative support, data entry, basic reporting, customer service, research, and content creation are increasingly being performed by software rather than people. Too many businesses are keen to replace several employees with an AI platform to reduce costs.

However, the long-term financial reality may prove more complicated. While AI is undoubtedly eliminating some entry-level jobs, there is growing evidence that the true cost of implementing, maintaining, governing, and scaling AI systems may eventually exceed the savings gained from reducing junior staffing. What appears economical today may become significantly more expensive tomorrow.

From a purely short-term perspective, the decision makes sense. If one AI subscription can automate work previously spread across multiple employees, savings appear immediate and measurable. Yet business owners often overlook a critical issue: AI itself is not a one-time investment. It introduces an entirely new category of ongoing costs that are far from predictable.

A common misconception is that AI operates like conventional software, where a fixed monthly fee grants unlimited utility. In reality, many AI platforms incur costs that scale directly with usage, making them fundamentally different from traditional applications. But these large language models (LLMs) used by AI require enormous computing resources and rely on vast data centres, specialist hardware, electricity consumption, network capacity, storage systems, and ongoing model updates. And as AI usage increases, expenses will increase rather than decrease. Regardless of whether a company develops its own models or relies on external AI providers, the substantial costs associated with computing power, data centres, networking, and maintenance are eventually reflected in the prices paid by customers. But unlike an employee, whose salary increases gradually over time, AI costs scale dramatically with usage.

Uber recently revealed just how quickly AI costs can spiral. According to the company’s technology leadership, Uber exhausted its entire 2026 budget for AI-assisted coding in only four months. By March, 84% of its engineers were using Claude Code to write software, and around 70% of new code updates were being generated with AI assistance. While adoption was widespread, the business value was less clear. Uber President and COO Andrew Macdonald acknowledged that soaring costs did not appear to translate directly into a proportional increase in customer-facing features or product improvements.

Uber is not alone. Microsoft, despite investing billions in OpenAI and reporting that AI generates a significant portion of its internal software, has reportedly scaled back the use of certain AI coding tools within parts of the organisation after costs became difficult to justify. Other companies have encountered even more dramatic surprises. Axios reported that one organisation accumulated a Claude-related bill of approximately $500 million in a single month after failing to implement effective usage limits. These incidents highlight a growing concern across the industry: while AI can dramatically increase output, the associated costs can rise just as quickly, and the relationship between spending and measurable business value is not always straightforward.

Current AI pricing reflects an intensely competitive market in which providers are spending vast amounts of capital to acquire customers and establish market dominance, while offering AI services at prices that may not fully reflect their long-term operating costs. Many AI companies continue to spend heavily on infrastructure, research, and development. As demand grows, the industry will face increasing pressure to recover these investments. Organisations that have become accustomed to low-cost AI services will see subscription fees rising and limits applied based on usage.

A more realistic scenario is one where AI transforms entry-level roles rather than eliminating them altogether. Rather, it is one where AI changes the nature of entry-level work. Instead of performing repetitive tasks manually, junior workers may focus on supervising AI systems by validating outputs and managing exceptions. Organisations that combine human talent with AI capabilities are likely to achieve better outcomes than those that rely exclusively on either.

All that said, we are starting to see some backtracking. Ford has recently hired back some human engineers after AI failed to match their skills and experience. In July 2025, the Commonwealth Bank of Australia axed 45 positions, stating: “Our investment in technology, including AI, is making it easier and faster for customers to get help, especially in our call centres.”  One month later they reversed that decision because call volumes rose at the bank, leading to increased overtime for the remaining staff and management being drafted to help answer phones.

AI will undoubtedly remain a transformative technology, but replacing entry-level workers is not the same as eliminating costs.