The Human Skills Becoming More Valuable as AI Improves

What happened

Two independent, large-scale workforce reports published within about a year of each other reached a similar conclusion. The World Economic Forum's Future of Jobs Report 2025, based on a survey of employers representing more than 14 million workers globally, found that alongside a jump in demand for AI and technical skills, "human" skills — creative thinking, resilience, flexibility, curiosity, leadership, and empathy and active listening — were rising in parallel, not shrinking. LinkedIn's Skills on the Rise 2026 report, built from hiring and skill-acquisition data on its own platform, found the same pattern from a different data source: skills like conflict mitigation, public speaking, stakeholder management, and cross-cultural communication are among the fastest-growing in demand, running at close to the same growth rate as technical AI skills like prompt engineering and AI implementation.

Why it matters

The common assumption is that AI improving means human skills matter less. Both of these reports point the other way: as baseline AI capability becomes something more employers have access to, the skills that differentiate one person or team from another shift toward the things AI still can't reliably do — reading a room, de-escalating a conflict, building trust across a negotiation, leading people through ambiguity.

What's verified, and what's still developing

What's solid: both reports are large-sample, methodologically transparent, and independently arrived at a similar conclusion using different data (WEF surveys employers directly; LinkedIn analyzes actual hiring and skill-acquisition behaviour on its platform). That kind of convergence across different methods is a meaningfully stronger signal than either report alone.

What's still uncertain: "skills on the rise" in job postings and hiring data measure what employers are asking for right now, not a guaranteed 10-year forecast. Labour-market demand can shift faster than any report captures, and these findings are strongest for the US and other markets with heavy LinkedIn/formal-hiring-market penetration — they may understate what's happening in informal or non-Western labour markets.

The pattern underneath both reports

Putting the two datasets side by side (something neither report does on its own) shows the same shape from two different angles:

The overlap is the finding: two organizations measuring different things (what employers say they want vs. what actually gets people hired) landed on the same shape — technical AI skills and human relational skills rising together, not one replacing the other.

What people commonly misunderstand

The mistake is treating this as "good news, don't worry about AI." That's not what the data says. AI and technical literacy are also rising fast in both reports — arguably faster in raw numbers. The finding isn't that human skills insulate you from needing technical fluency. It's that once a baseline level of AI literacy becomes common (which both reports suggest is happening quickly), the differentiation between candidates and employees increasingly comes from the human layer on top of it.

An original framework: The AI-Resistant Skill Check

Before assuming a skill is "safe" from AI, it's worth running it through a short check rather than a hunch. Ask, honestly, of any specific skill in your own work:

1. Does it require reading a specific person's emotional state in the moment, not just processing what they said? (Empathy, de-escalation, negotiation reading — currently hard for AI.)
2. Does doing it well depend on people trusting you specifically, not just trusting that the output is correct? (Leadership, stakeholder management, client relationships.)
3. Does it require taking responsibility for a judgment call under real uncertainty, where being wrong has a cost you personally bear? (Accountability-bearing decisions, not just decision-support.)
4. Would the task change meaningfully if the physical or social context changed unpredictably? (Adaptability in the room, not adaptability on paper.)

A skill that scores "yes" on two or more of these is the kind both reports show rising. A skill that's mostly pattern-matching against existing information, with a single clearly correct answer and no relationship or accountability component, is the kind more exposed to AI substitution — which is exactly the kind of task both reports show organizations already automating first.

What this means for an ordinary reader

If you're deciding where to invest a limited number of hours this year, both reports point toward the same allocation: don't skip AI literacy, but don't stop there either. Pair it with a deliberately developed human skill — the kind that shows up in the tables above, and the kind that requires you, specifically, showing up in a room or a difficult conversation. That combination is what both reports call the most competitive profile, not either skill set alone.

Next steps

Our Decision Matrix tool can help you weigh where to actually spend limited learning time between a technical AI skill and a human skill, based on what matters most in your specific role — rather than defaulting to whichever headline sounds more urgent this month.