Technical Context: I Look at the Signal, Not the Hype
I reviewed Anthropic's study Labor Market Impacts of AI: A New Measure and Early Evidence, published in early March 2026, and to me, its value lies not in loud conclusions, but in the methodology. The team didn't stop at the theoretical question of "what LLMs can automate," but combined the potential applicability of models with actual automation usage on Anthropic's platforms.
This is exactly what makes this paper stronger than most public reviews. I constantly see the same mistake in client projects: people confuse technical feasibility with the real integration of artificial intelligence into the operational workflow. Anthropic precisely highlights the gap between the ability to automate a task and the fact that a company has actually embedded AI into its process.
I interpret the key figures as follows. Since late 2022, there has been no systemic rise in unemployment in the most AI-exposed professions, but a hiring slowdown signal is noticeable for the 22–25 age group: the probability of finding a job decreased by about 14% relative to the baseline, although statistical significance is borderline there. For me, this is not a final verdict, but an early indicator of a shift in entering the profession.
I also paid attention to the exposure distribution. It's not just routine roles like data entry taking a hit, but also intellectual functions: programming, customer service, and financial analysis. This aligns perfectly with what I see in practice: LLMs primarily transform not "simple work," but work involving text, code, regulations, reports, and interfaces.
Impact on Business and Automation: The Most Organized Win, Not the Boldest
I don't conclude from this research that "people will be replaced." I draw a different conclusion: companies are already reaping the economic benefits of AI automation without a sharp increase in layoffs because they are cutting back on new hiring, task completion times, and the share of manual labor in white-collar processes rather than slashing headcount.
The winners will be those who restructure their role architecture before others. If previously a junior was a cheap way to cover volume, now part of that volume is covered by LLMs, API integrations, and agentic scenarios. This means businesses don't just need executors, but employees who can formulate a task, validate model results, and manage a hybrid human-AI loop.
Companies that implement AI as a set of isolated chatbots will lose. Such an approach lacks controllability, metrics, and a new operational model. In my experience at Nahornyi AI Lab, real AI automation begins when we rebuild the entire process: incoming requests, routing, quality assurance, CRM, ERP, internal knowledge bases, access rights, and auditing model actions.
For HR and operations directors, this presents a clear crossroad. Either you continue hiring based on the old role matrix and end up with an excessive layer of manual operations in a year, or you transition to an architecture where AI solutions for business are embedded in teams' daily work, and hiring focuses on supervision, exception handling, and domain expertise.
Strategic View: Shortages Will Arise Where Least Expected
I consider the most underestimated finding to be not the slowdown in hiring young talent itself, but where the new shortage will emerge. The market has long debated models replacing humans, but in practice, the most valuable specialists will soon be those capable of designing a working integration of LLMs, APIs, company data, and control points.
This is exactly why AI solution development and AI architecture are becoming a managerial necessity rather than an IT novelty. In my projects, the client's most valuable employees are no longer those who produce text or reports the fastest, but those who can decompose a business function into automatable steps, set escalation rules, and measure the impact post-launch.
Anthropic isn't recording a spike in unemployment yet, and I wouldn't stretch these data to predict an apocalypse. But I certainly wouldn't ignore this early pattern: the market is adopting LLMs through a slowdown in profession entry, not through immediate mass layoffs. For businesses, this is an even more serious scenario because it alters the talent pipeline, not just the current payroll.
This analysis was prepared by Vadym Nahornyi — Lead AI Expert at Nahornyi AI Lab, specializing in AI automation and practical business architecture. If you want to do more than just discuss the news and actually translate it into a clear action plan for your company, I invite you to a substantive conversation with my team at Nahornyi AI Lab: we will analyze where AI implementation will pay off for you, which roles will change first, and how to execute AI integration without chaos.