What new AI labor market research means for college-bound families.

If you've been wondering whether the degree your child is pursuing will still matter in five years, you're not alone.

Two major research reports dropped recently that every college-planning family should know about. One from Anthropic (the company behind the AI system Claude), studying how AI is affecting employment right now. The other from Lightcast, a labor market data firm, analyzing how job skill requirements are shifting across the entire U.S. economy.

The Number That Should Stop Every Parent in Their Tracks

According to Lightcast's analysis of millions of U.S. job postings, the average job's skills have changed 33% from 2022 to 2025.

In just three years, the actual skill requirements for roles across the economy have been reconfigured by roughly a third, and the pace is accelerating.

This means a degree or skill set your child begins building today will look materially different by the time they graduate.

What the AI Research Found

Anthropic's study adds an important layer: AI is theoretically capable of handling 94% of tasks in computer and math fields, but in practice, it's only actually doing about 33% of them.

That gap matters. It tells us capability and deployment are two different things, and that there's still a window of time to prepare thoughtfully rather than reactively.

The research also found something that surprised many people: Since the release of ChatGPT, there's been no dramatic wave of unemployment among people in AI-adjacent roles. But there is early evidence of hiring slowdowns for younger workers entering those fields.

Translation: the ground is shifting under the entry point to these careers. For students about to launch into college or the workforce, it means they may need to get more creative about gaining real experience in the domains they care about.

Which Skills Are Actually Disappearing

This is where the Lightcast data gets very specific, and very useful for families making education decisions.

The fastest-declining skill categories in U.S. job postings between 2022 and 2025 include JavaScript and jQuery (down 60%), Graphic and Visual Design (down 59%), Web Design and Development (down 56%), and Java (down 57%). Coding languages. Design execution. Content production tools. Precisely the skills many parents assume are a safe bet in a tech-driven world.

The pattern makes sense when you understand what AI is good at: it excels at discrete, repeatable, output-focused tasks. Writing a block of code. Generating a layout. Producing a first draft. These are the skills it's absorbing fastest.

Skills tied to routine implementation are declining, while higher-order skills like system design, strategic thinking, creative direction, and ethical oversight are becoming more important.

Durable Skill Building

With both data sets in mind, some clear patterns emerge about what skills have staying power in an AI world.

Critical thinking and judgment. AI can generate options. It can't weigh them against your values, your client's needs, or the full complexity of a situation. Students who learn to think through ambiguity, not just produce answers, will be the ones directing AI, not being replaced by it.

Communication and relational intelligence. The ability to read a room, build trust, navigate conflict, and inspire people is deeply human. Every field that requires working with or for people will continue to reward it.

Adaptability. A job market where one-third of skill requirements shift in three years doesn't reward students who know the right answers, it rewards students who know how to learn new ones. Students who develop a growth mindset and understand how they learn will be better positioned.

Domain expertise + AI fluency. AI literacy is no longer a specialized, technical requirement. It's becoming a baseline expectation across the entire labor market, but it doesn't mean machine learning is the only viable degree path. This is where a skill stacking approach becomes powerful. Skill stacking is the strategy of combining multiple, distinct, and often unrelated skills to create a unique, highly valuable, and hard-to-replicate professional profile, rather than specializing in just one area.

What This Means for College ROI and The Future of Work

Is college still worth it? The honest answer: it depends on what you're optimizing for.

A degree that trains your child primarily to execute tasks that AI can now handle cheaper and faster? Probably not worth it. But a college experience that builds judgment, deepens domain expertise, develops relationships, and forces a young person to grapple with hard problems in community with others? That's still extremely valuable, but it's not a given at every university.

The question to ask isn't just "what will my child major in?" It's "what kind of thinker will they become, and what real experiences will shape that?"

That's the conversation Traverse exists to help families have — early enough to matter.

Traverse helps students and families navigate the college and career transition with clarity, strategy, and confidence. Curious about how we think about AI and future-readiness in our programs? [Get in touch.]

Sources: Anthropic, "Labor Market Impacts of Claude" (2026); Lightcast, "The Speed of Skill Change: 2026 Update"