Imagine you run a clinic and discover that an artificial intelligence tool can perform much of a new employee's administrative workload. It organizes documents, reviews information, and prepares reports. Why hire someone who is just starting to learn?
That decision, which might seem logical to any business owner, is already reshaping hiring practices at one of the tech sector's most prominent companies.
Anthropic, the developer behind the Claude artificial intelligence system, revealed that it is bringing on far fewer entry-level workers across technical areas. Since early 2025, the share of these positions decreased from 21% to 14%. Among new hires, the drop was even steeper: from 27% to 9%.
The reason lies in what its AI systems can currently perform.
By August 2026, Anthropic's artificial intelligence agents had accumulated six times more operating hours than its human staff. These systems now handle tasks such as coding and bug detection that previously called for greater human involvement.
The internal results are striking. The engineering team multiplied its code deliveries by 5.1. The legal department managed 3.1 times more contracts. The sales division finalized 1.7 times more deals, and accounting handled 1.4 times more invoices.
However, there is an important caveat: Anthropic creates artificial intelligence and is uniquely prepared to use it. Its results do not necessarily reflect what will happen in other sectors. In fact, no measurable productivity boosts from AI have been observed across the general economy yet.
Why it matters
The issue goes beyond how many tasks a machine can resolve. It centers on what happens when companies stop hiring people to learn how to do them.
Consider an auto repair shop. An assistant starts with simple tasks, observing experienced technicians and making mistakes under supervision. Over the years, that assistant can become the expert who solves the toughest problems.
Now imagine that assistant position vanishes because an automated tool handles those initial responsibilities.
The business might gain efficiency today, but where will its specialists come from tomorrow?
Researcher Nolan Lovett has studied this issue. His premise is that individual companies can benefit from automating basic positions, while the cost of training fewer professionals ultimately affects everyone.
According to his analysis, the consequences could become visible between 2030 and 2045, especially in fields like programming, financial analysis, and legal research.
This is a projection, not a certainty. It is also possible that AI will deeply transform the professions we know today.
What it means for your business
The lesson is not to avoid artificial intelligence or keep hiring people for tasks that a tool can perform better.
It is about distinguishing between automating labor and removing learning opportunities.
At a real estate agency, for example, AI could draft property descriptions while an entry-level agent learns to negotiate and assist clients. In a restaurant, it could help with administrative reports while a new manager develops decision-making judgment.
Technology can remove repetitive work without taking away your team's opportunity to gain experience.
This week, choose one task your company already handles with AI, or one you could automate. Identify what skills the responsible person previously developed, and decide how you will keep teaching them.
Because saving time is important. But ensuring that someone knows how to run the business tomorrow is equally vital.