Is AI Coming for Tech Jobs? What Anthropic’s Latest Research Reveals
Artificial intelligence is already changing AI and tech jobs, from software development and IT to the way companies hire and evaluate technical talent. But when it comes to its actual impact on jobs, the reality has been harder to pin down. Recently that picture is starting to change.
Anthropic has continued expanding its research into how AI is being used at work, moving beyond theoretical predictions about which jobs could be automated and looking at what people are actually doing with AI.
The latest research points to a more complicated picture than either extreme suggests.
AI isn't simply replacing workers. But it is changing the way work gets done, and in some areas of technology, that change is happening quickly.
AI Adoption Is Moving Quickly From Experiment to Workflow
Early AI adoption was relatively simple: ask a chatbot a question, generate some text or use it to help with a specific task. That's changing.
Anthropic's latest Economic Index research shows that AI sessions are increasingly becoming longer-running, more autonomous workflows through tools such as Claude Code and Cowork. Instead of simply assisting with an individual task, AI can now participate in a much larger portion of a project.
The question is no longer just whether AI can perform a task. It's increasingly whether a company can reorganize a workflow around AI. And that's a much bigger question for employers.
Which Tech Jobs Are Most Exposed to AI?
AI exposure remains concentrated in work that is digital, structured and largely performed on a computer. Software development remains one of the clearest examples.
Anthropic's research has found that computer and mathematical occupations account for a disproportionately large share of AI usage compared with their share of the overall U.S. workforce.
Other areas receiving significant AI exposure include:
Computer programming
Data analysis
Customer service
Financial analysis
Sales and outreach
Technical writing
Information technology
But exposure does not automatically mean elimination. A job can be highly exposed to AI while still requiring a human to make decisions, verify results, manage risk and understand the larger business context.
That's one of the most important distinctions in the AI employment conversation.
A task can be automated without an entire job disappearing.
The Bigger Shift: AI Is Taking On More of the Work
Anthropic's newest survey data provides a more striking look at where this is headed.
Nearly six in 10 respondents said they expect AI to be capable of handling a larger share of their work over the following 12 months. More than one-third expect AI to be able to perform most or nearly all of their work tasks.
That doesn't mean one-third of jobs are about to disappear. It does suggest something potentially more consequential for employers: the amount of human labor required to produce a given amount of work may be changing.
For companies, that could affect how teams are structured, how many people are needed for certain projects and what skills employers prioritize when they do hire.
How AI Is Changing Software Development Jobs
Perhaps nowhere is this shift more visible than software development. Anthropic's 2026 research analyzed roughly 400,000 Claude Code sessions between October 2025 and April 2026. One of the clearest findings was how the division of labor is changing.
In a typical coding session, people tend to make most of the planning decisions while Claude increasingly handles the execution decisions, or how to accomplish it. That's a significant change in the definition of technical work.
The developer is increasingly becoming something closer to an architect, director and reviewer of AI-assisted work. And that could change what employers look for when hiring.
A candidate who simply knows how to write code may no longer have the same advantage as someone who understands the system, can define the problem, can direct AI effectively and can recognize when the resulting code is wrong.
The technical fundamentals still matter. But the ability to work effectively with AI may increasingly matter alongside them.
Expertise May Actually Become More Valuable
There is an interesting wrinkle in Anthropic's coding research. The more domain expertise a person brings to a Claude Code session, the more work Claude tends to perform per instruction. That's important because it challenges the idea that AI simply makes experience irrelevant. In practice, the opposite may be happening.
An experienced developer who understands architecture, security, business requirements and potential failure points can give AI better direction and evaluate its output more effectively.
The technology can accelerate the work. But someone still needs to know what good work looks like. That creates an interesting possibility for the IT labor market: AI may increase the value of technical expertise even as it reduces the amount of routine technical work humans need to perform.
Why AI Skills Are Becoming More Important to Employers
Another finding deserves attention. Anthropic's research suggests that people who use AI more extensively report significant productivity gains.
Among surveyed users, 86% reported gains in the speed of their work, 82% reported gains in scope, meaning they could accomplish more or take on new tasks, and 69% reported gains in quality.
More than half also said AI had made their skills more valuable.
These are self-reported results, so they should not be treated as proof that AI improves every worker's performance. But they point toward an emerging hiring consideration.
It may no longer be enough for an IT candidate to say, "I've used AI."
Employers may increasingly want to know:
Which AI tools have you used?
How have you incorporated them into your workflow?
How do you evaluate AI-generated output?
How do you identify errors or security issues?
Can you use AI without losing control of the underlying process?
Can you explain when AI should, and should not, be used?
That's a very different definition of AI experience.
Early-Career Workers Could Face a Tougher Starting Point
This may be one of the most important issues for recruiters.
Anthropic's research shows that early-career workers report that AI can perform a larger share of their work and express greater concern about job displacement. That creates a difficult question for the technology industry.
Historically, junior employees have learned by handling the routine work that surrounds more complex assignments. But what happens when AI can handle much of that routine work?
Companies may need fewer people to perform entry-level tasks—or they may need to rethink what an entry-level position actually looks like. At the same time, organizations still need people who understand the fundamentals behind the technology.
The challenge may be shifting from getting experience by doing routine work to getting experience by learning how to manage, evaluate and improve AI-assisted work.
For recruiters, that could make internships, apprenticeships, mentoring and structured training more important.
AI Isn't Affecting Everyone the Same Way
The impact of AI is also unlikely to be evenly distributed across the workforce.
Anthropic's research shows differences in how people use AI depending on their occupation, experience and level of automation. It also found that more experienced workers tend to estimate that AI can perform a smaller share of their work than less-experienced workers.
One possible explanation is that experienced professionals have accumulated more tacit knowledge, the context, judgment and relationships that aren't easily captured in a job description.
That's particularly relevant in technology. A senior cybersecurity professional, for example, isn't simply executing a checklist. They may be assessing risk, interpreting incomplete information, understanding an organization's vulnerabilities and deciding what deserves immediate attention.
AI can assist with those tasks. But the human still needs to understand the consequences of getting them wrong.
Is AI Really Replacing Tech Jobs?
The evidence still doesn't support a simple answer.
Anthropic's research has not shown a broad collapse in employment among workers in highly AI-exposed occupations. But that doesn't mean the labor market is standing still.
One of the biggest changes may happen before mass layoffs ever become visible.
It may be how companies decide when to hire, and how many people they need.
If an experienced employee can accomplish substantially more with AI, a company may decide to expand an existing team less aggressively.
That doesn't necessarily mean fewer jobs overall. It could mean fewer people doing certain routine tasks while demand grows for people who can manage more complex systems and AI-assisted workflows.
For recruiters, that's a critical distinction. The question isn't simply whether AI will eliminate jobs. It's whether AI will change the composition of the workforce.
What AI Means for Tech Talent and Hiring
The technology labor market may be moving toward a different definition of expertise.
Technical professionals still need strong fundamentals.
But increasingly, employers may also value people who can:
Work effectively with AI coding and productivity tools
Evaluate AI-generated work
Identify errors and security risks
Design AI-assisted workflows
Make decisions AI cannot reliably make
Communicate technical information clearly
Understand the business problem behind the technology
Know when not to use AI
In other words, AI may not eliminate the need for technical talent.
It may raise the expectations placed on that talent.
The strongest candidates may increasingly be those who understand both the technology itself and how to use AI to extend their capabilities.
How AI Is Changing the Way Companies Hire
For years, employers asked: "Can this person do the job?"
AI may add another question: "Can this person use AI to do the job better?"
Anthropic's latest research suggests AI is moving steadily from an experimental tool toward something embedded in everyday workflows. And as AI becomes more capable, the distinction between someone who has simply experimented with AI and someone who knows how to integrate it into a professional workflow may become increasingly important.
That could affect everything from job descriptions to interviews to the way recruiters evaluate technical candidates. The technology is still evolving too quickly to know exactly what the workforce will look like five years from now.
But one thing is becoming clearer. AI may not replace the people who know how to do the work. It may increasingly change the value of people who know how to work with AI.
For technology employers and candidates alike, that's a shift worth paying attention to.
Looking for your next IT or cybersecurity role? Reach out to the team at Rekruitd. We specialize in connecting top-tier talent with the right opportunities.
