Workers with AI skills earn 56% more than their peers. The premium more than doubled in a single year. Something significant is happening in the labour market.
AI skills now outperform a postgraduate degree in terms of immediate labour market returns. The wage premium has more than doubled in a single year. Job postings requiring AI skills grew 7.5% while total postings fell 11.3%. The professionals who understand this and act on it will have a measurable edge within 12 months.
PwC publishes something called the AI Jobs Barometer. It’s a serious piece of research, based on analysis of close to a billion job ads from six continents, covering 15 countries and salary data across industries. The 2025 edition landed with a finding that stopped a lot of HR directors and L&D professionals in their tracks.
Workers with AI skills earn, on average, 56% more than their peers. One year before that, the premium was 25%. It more than doubled in twelve months. The report is explicit: this premium shows up in every industry analysed, not just technology roles.
Let that sit for a moment. The wage advantage for AI literacy more than doubled in a single year, and it’s reflected in actual job postings and actual salaries, not forecasts or projections.
This is not a fringe finding from a tech publication with a vested interest in AI adoption. PwC is one of the largest professional services firms in the world. This is their analysis of real hiring data. It reflects what employers are actually willing to pay.
AI skills now outperform a postgraduate degree in immediate labour market returns. That’s not a prediction anymore. It’s what the hiring data shows right now.
Twelve months ago, the premium was 25%. Now it’s 56%. Why has it nearly doubled in a year?
A few things happened simultaneously. First, the capabilities of AI tools took a significant step forward in 2025. The gap between what an AI-proficient professional can accomplish and what an AI-naive professional can accomplish got noticeably wider. Employers noticed.
Second, demand for AI skills started outpacing supply in a way that’s becoming structurally significant. Job postings requiring AI skills grew 7.5% over the past year. Total job postings fell 11.3%. That’s a scissors chart that puts pressure on wages for the people who have what employers need.
Third, the skills themselves are changing faster than most people can track. PwC’s research found that in roles highly exposed to AI, the skills employers ask for are changing 66% faster than they were a year ago. That acceleration makes it hard for even motivated professionals to keep up, which means the ones who do stay current command a premium for their currency.
The fourth factor is more subtle. Productivity data is starting to show up clearly. Industries highly exposed to AI have seen productivity growth nearly quadruple since AI tools went mainstream in 2022.[1] When productivity grows that much, the workers driving it capture more of the gains.
Not all AI skills carry the same earning power. PwC’s research found the premium applies “in every industry analysed,” but the specific value of individual skills varies significantly by role and field.
At the technical end of the spectrum, skills like machine learning engineering, data science, and AI model development command the highest premiums. These are real, but they require substantial investment to develop and are most relevant to people whose job is building AI systems. That’s not most professionals reading this.
At the more accessible end, the skills that show up consistently in non-technical roles are: effective use of AI tools (knowing how to prompt them, evaluate their outputs, and integrate them into real workflows), AI judgment (understanding what AI is good at and where it falls short), and AI communication (being able to speak clearly about AI capabilities in cross-functional conversations). PwC specifically calls out skills like prompt engineering as part of the AI skill set driving wage growth.
A marketing director who understands prompt engineering, can evaluate AI output critically, can integrate AI tools into campaign workflows, and can brief technical teams intelligently doesn’t just earn more. She gets better results, operates faster, and has credibility in conversations that non-AI-literate peers can’t enter.
For most business professionals, the practical skill set to focus on breaks down into three layers:
Effective use: Knowing how to prompt AI tools well, evaluate their outputs, and integrate them into daily workflows. This is foundational and learnable in weeks, not years.
AI judgment: Understanding what AI is good at and where it falls short. Knowing when to trust it, when to verify, and when to keep humans in the loop. This comes from consistent use and honest reflection.
AI communication: Being able to speak clearly about AI capabilities and limitations in cross-functional contexts. This is increasingly a leadership skill, and it separates people who use AI from people who can lead AI-informed teams.
One of the more counterintuitive findings from the PwC research is this: jobs are growing in virtually every type of AI-exposed occupation, including the ones most likely to be automated. That sounds paradoxical, but it makes sense when you think it through.
Automation doesn’t eliminate roles wholesale. It reshapes them. The HR director whose team used to spend 40 hours a week screening applications doesn’t get a team of zero. She gets a smaller team that’s more senior, because the entry-level screening work is handled by AI, and the remaining human work is higher-judgment hiring decisions, candidate experience management, and employer brand strategy.
The IMF published research in January 2026 showing this pattern across 30 countries. AI exposure correlates with job growth, not job loss, when workers have the skills to shift their focus to the higher-value parts of their roles.[2]
This plays out in legal, finance, marketing, operations, consulting, real estate, and education. Every professional field has AI-exposed roles, and the pattern is consistent: AI-literate practitioners are being promoted into expanded roles, while those without AI skills are watching their responsibilities narrow.
Here’s the part that’s genuinely hard to explain. Despite a 56% wage premium being documented and widely reported, roughly 70% of workers aren’t actively developing AI skills.
The most common reasons people give:
“I don’t know where to start.” This is the most defensible reason, and it’s getting less true every month. But the landscape of courses, tools, and credentials is genuinely confusing, and not everyone has a trusted guide.
“My job doesn’t really involve AI yet.” This is usually incorrect, and the people who believe it most firmly are often the ones most at risk. If your job involves reading, writing, research, data analysis, customer communication, or decision-making, AI already applies to what you do.
“I don’t have time.” This is real, but it’s also a bit circular. The time investment required to reach functional AI literacy is probably 10-15 hours of focused learning and practice. That’s one intensive weekend or two hours a week for a month. The return on that time, measured in both compensation and capability, is significant.
“I’m worried about my job.” This one is understandable emotionally, but it reflects a misreading of the evidence. The data consistently shows that AI-literate workers in AI-exposed fields are more secure, not less.
A practical path for a business professional who isn’t technical and doesn’t want to become a data scientist.
Pick one AI tool you’ll use every day and commit to using it seriously for four weeks. ChatGPT, Claude, and Gemini are all solid choices. The goal isn’t to pick the “best” one. It’s to build genuine familiarity with how these systems respond to different types of prompts and where they require correction.
During this month, focus on your actual work. Don’t practice on abstract exercises. Use AI to draft emails, summarize long documents, prepare for meetings, and do research. Pay attention to where it helps and where it goes wrong. Keep a short log of observations.
If you want a more structured starting point, our guide on building a custom GPT for your specific workflow is a practical, non-technical way to develop this foundational layer.
Add one tool that’s specific to your function. If you’re in HR, look at AI-assisted recruiting platforms like HireVue or LinkedIn Recruiter’s AI features. If you’re in finance, explore AI features in your existing FP&A software. If you’re in marketing, experiment with AI in your content or analytics stack. Function-specific AI literacy is particularly valuable because it’s harder to replicate and more directly tied to your work outcomes.
Read one article a week from a credible source about how AI is affecting your industry. Not to stay current on every tool, but to develop the vocabulary and perspective that lets you participate in conversations about AI strategy. This is the layer that moves you from “user” to “informed practitioner,” and it’s the layer that shows up in performance reviews and promotion conversations.
For a broader understanding of where AI models stand right now and what they can do for basic context on what the tools can do now, our breakdown of the latest ChatGPT for busy professionals gives useful context for this stage.
If you manage people, the 56% wage premium has implications beyond your own career. It tells you something about where the talent market is going and what your team needs to stay competitive.
Organizations that started AI upskilling programs in 2023 and 2024 are now seeing measurable results. They’re able to do more with the same headcount, their teams are more engaged (because people generally like getting better at things that matter), and they’re better positioned to use new AI capabilities as they emerge.
The most effective upskilling programs share a few characteristics. They’re practical rather than theoretical, tied to actual work rather than generic AI courses. They give employees time and permission to experiment. They create internal communities where people can share what’s working. And they’re supported by leaders who model the behavior, which means leaders who actually use AI tools themselves, not just sponsor the training budget.
If your organization doesn’t have a formal program, starting informally is fine. Identle who are already using AI well and make them visible. Share practical examples of what’s working. Create a low-stakes channel where people can ask questions without embarrassment. The culture shift matters as much as the curriculum.
No. The wage premium data covers a wide range of AI competencies, including non-technical ones like prompt engineering, AI-assisted analysis, and effective use of AI tools in professional workflows. Coding is valuable for people who want to build AI systems, but most business professionals don’t need it to significantly benefit from AI literacy.
PwC’s research covers 15 countries and multiple industries, so it reflects a broad average rather than one sector. The premium varies by role and field. Technology roles tend to have higher premiums for technical AI skills, while the premium for general AI literacy is more consistent across industries including finance, healthcare, legal, marketing, and operations.
The evidence suggests that the premium for AI skills is more likely to stay high than disappear. As AI capabilities grow, the ceiling for what skilled AI users can accomplish also rises. The pattern in previous technology waves (internet, mobile, cloud) is that the advantage shifts from knowing the technology to knowing how to apply it strategically. That’s a human skill, and it compounds over time.
For functional proficiency with AI tools in a professional context, most people reach a genuinely useful level in 4-8 weeks of consistent use. That’s not expertise, but it’s enough to be noticeably more effective. Deeper skills that show up in more technical roles take longer, typically 6-12 months of deliberate practice and learning.
Yes. Employers increasingly don’t differentiate between formal training and self-taught proficiency because the outcomes are what matter. Be specific about which tools you use, how you use them, and what you’ve accomplished with them. “Proficient in ChatGPT” is weaker than “Used AI tools to reduce weekly reporting time by 60% and produce first drafts of client-facing analysis.” Specifics are more credible than labels.
This article was researched and written by Sana for Future Factors AI. Primary source: PwC 2025 Global AI Jobs Barometer (published June 2025, based on analysis of close to a billion job ads from six continents). Additional sources include IMF research on AI and labour markets, Gloat’s AI Workforce Trends analysis, and World Economic Forum reporting on AI wages and job quality.
Weekly tips. Real examples. Practical help for busy professionals.
We care about your data, check out our privacy policy.