A Number Making People Reconsider Everything
Here's a number that's making a lot of people reconsider their career paths: $18,000.
That's the average annual salary premium for workers with AI skills on their resume, according to PwC's 2025 Global AI Jobs Barometer. Workers who can demonstrate AI proficiency earn about 28% more than their peers in comparable roles.
For specific AI skills, the premiums are even more dramatic. Prompt engineering commands a 56% wage premium — up from 25% just a year earlier. Machine learning specialists earn 40% more. Deep learning adds 27%. Natural language processing adds 19%.
At first glance, the conclusion seems obvious: drop everything, learn AI skills, cash in.
Not so fast. Because the story is a lot more complicated — and a lot more interesting — than "AI skills = more money."
The Paradox: High Wages, Uncertain Volume
Here's the part the "learn AI!" crowd glosses over.
While AI skills command huge wage premiums, they haven't created net employment growth in AI-exposed occupations. Unlike other emerging skills that both raised wages and created more jobs, AI skills are in a weird position: they're worth a lot per worker, but there aren't necessarily more workers needed.
What does this mean practically? It means AI skills are like investment banking skills in the 2000s: fantastic pay for those who get in, but the number of seats at the table isn't growing proportionally.
The hiring numbers tell the story: AI/ML hiring grew 88% year-over-year. Sounds great! But administrative role hiring fell 35.5%, and entry-level hiring fell 73.4%. AI is creating high-paying jobs at the top while removing lower-paying jobs at the bottom.
So the question isn't just "should I learn AI skills?" It's "what kind of AI skills, combined with what domain expertise, will actually make me more valuable?"
Three Career Strategies (Pick Yours)
Based on the wage and employment data, I see three viable strategies. None of them is "become an AI engineer." They're more nuanced than that.
Strategy 1: Domain Expert + AI Fluency
This is the highest-value play for most professionals. Take your existing domain expertise — accounting, marketing, operations, law, medicine, whatever — and layer AI fluency on top.
The data supports this: the biggest productivity gains from AI come in domain-specific applications. A lawyer who can use AI to accelerate legal research while applying superior legal judgment. A financial analyst who uses AI to build models in minutes while bringing deep market understanding to the interpretation. A marketer who uses AI to generate and test campaign variations while understanding the brand strategy that makes certain variations better than others.
This strategy works because it creates complementarity — AI handles the parts you're less good at (or less interested in), while you focus on the parts AI can't match. And the premium for this combination is real: you're not competing with AI. You're multiplied by it.
Strategy 2: AI-Native Builder
This is the pure technology play. Learn machine learning, deep learning, AI systems architecture. Build AI products. Become the person who creates the tools that everyone else uses.
The premiums here are highest ($160,000 median salary for AI professionals), but so is the competition, and the pace of change is brutally fast. The specific frameworks and tools you learn today may be obsolete in 18 months. This strategy requires continuous learning at a pace most people find exhausting.
It's the right play if you're technical, love building things, and can tolerate constant skill renewal. It's the wrong play if you're hoping to "learn AI once" and coast.
Strategy 3: Human-Skill Specialist
This is the contrarian play, and I think it's underrated.
As AI handles more cognitive work, demand for deeply human skills is increasing. Leadership, coaching, negotiation, complex client relationships, ethical reasoning, creative vision, team building, conflict resolution.
David Deming's research at Harvard shows that jobs requiring high social skills grew by 12 percentage points over three decades while socially isolated cognitive jobs shrank. McKinsey projects social and emotional skill demand rising 11-14% by 2030.
The logic is straightforward: as AI handles "thinking" work, humans increasingly handle "feeling" work. The therapist, the executive coach, the enterprise sales leader, the creative director, the community builder — these roles become more valuable, not less, as AI handles the analytical parts of work.
This strategy won't command the eye-popping premiums of Strategy 2. But it offers stability, meaningful work, and growing demand — which, for many people, is a better career deal than a high salary on an uncertain timeline.
The Real Answer
So should you pivot to AI skills?
My honest answer: it depends on who you are and where you're starting from.
If you're already technical and enjoy building things, yes — deepening your AI expertise is a no-brainer. The premiums are real and the opportunities are expanding.
If you're a domain expert in any field, don't pivot — augment. Keep your domain expertise and add AI fluency. This combination is the sweet spot where demand exceeds supply and premiums are highest relative to the investment required.
If you're naturally drawn to human connection, leadership, and interpersonal work, don't feel pressured to become technical. The world needs more great leaders, coaches, and relationship-builders — not fewer. AI makes your skills more valuable, not less, by handling the analytical work that used to compete for your time.
The $18,000 premium is real. But it's not a universal invitation to "learn AI." It's a signal to figure out your specific version of AI fluency — the version that amplifies your unique strengths rather than turning you into a generic AI worker competing with every other generic AI worker.
Because in the end, the biggest premium doesn't go to people who know AI. It goes to people who know something else really well and can orchestrate AI to amplify that something else.
Your domain expertise isn't a liability in the AI age. It's your ticket.