Anthropic: AI is not eliminating jobs yet, but the gap between users is already growing

Anthropic has published the first data on AI's impact on employment—and they demolish both of the most popular narratives at once. The company found that mass displacement of workers is not occurring. Employees are not losing their jobs to chatbots and language models—at least not on the scale that analysts feared. But simultaneously, something else emerged, and this something is significantly more important: the gap between those who can work with AI and those who cannot is growing rapidly. Power users—people with a high level of tool mastery—are pulling ahead in productivity, work quality, and task completion speed. And this gap accumulates every single day.

The research is based on real data about Claude usage and other Anthropic AI tools. The sample consists predominantly of knowledge workers—analysts, developers, marketers, editors, managers. The pattern is clear: those who invested time in mastering AI tools, learned to formulate tasks correctly, and built workflows around them, receive disproportionately larger benefits from the same technologies available to everyone else.

According to early Anthropic data, the gap doesn't disappear over time—it grows. The longer an experienced user works with the tool, the greater their lead over a colleague just starting to understand basic features.

At first glance, "AI is not destroying jobs right now" sounds like good news. And in the short term, it is. Companies aren't laying off people in bulk, the HR apocalypse is postponed indefinitely. But the problem is different: if within a single profession, company, or sector there emerges a multiple difference in productivity between those who can use AI and those who cannot—this creates an entirely new type of inequality. Not between professions—as programmers and cashiers were once contrasted—but within a single profession.

Two marketers with identical experience, education, and salary. One mastered Claude, learned to generate and iterate briefs, analyze data, build content plans in minutes. The other didn't. After a year, their value to the employer is incomparable, even if their positions and titles haven't changed.

What exactly distinguishes power users is not a technical background, as you might think. The gap is created by specific behavior. Experienced users don't perceive AI as an advanced search engine or draft-generation tool: they build dialogues, clarify, iterate, provide context upfront, and manage the model as a thinking partner. They embed AI at every stage of work—from problem formulation to final result verification, not just at the "write text" stage. They experiment, record successful patterns, and build personal prompt libraries. They treat the tool like a junior employee—give clear instructions, check output, provide feedback.

It is precisely this gap in approach that creates the gap in results that Anthropic documented.

For business, the conclusion is direct: investment in employees' AI literacy is not a nice HR presentation bonus, but a measurable competitive factor. A company where 20–30% of employees can work with AI at a power-user level gains asymmetric advantage over a competitor where such people are 5–10%.

For the labor market, the signal is more complex. If the gap is widening within professions, it changes hiring logic, career advancement, and compensation. AI literacy is rapidly becoming as basic a requirement as computer literacy or office software proficiency once became—and those who lag in this transition risk finding themselves in a structurally disadvantageous position regardless of their other competencies.

Anthhropic directly discusses the risks of future displacement and widening workforce gaps. Reading between the lines: the company building the most powerful AI systems is warning us—we are not in the replacement phase. We are in the differentiation phase. And this may be a more insidious phase because it happens invisibly, without headlines about layoffs and waves of technological unemployment. It happens daily, in every office, in every meeting, in every project—as a growing gap between those who managed and those who are waiting.