We analyzed 289 jobs across 26 industries to determine which roles face the highest automation risk. Here's what the data reveals about the future of work.
We analyzed 289 jobs across 26 industries to determine which roles face the highest automation risk. Here's what the data reveals about the future of work.
Source: What About AI? — James Perkins
In January 2025, we set out to answer a question that was keeping professionals across every industry up at night: Is my job safe from AI? Not with speculation or hot takes—but with data. We built a database of 289 professions across 26 industries, scored each one on displacement risk, and what we found surprised even us.
Twelve months later, the acceleration has only confirmed what our numbers predicted. McKinsey Global Institute's 2024 report estimates that roughly 30% of hours worked in the United States could be automated by 2030. Goldman Sachs puts it even more starkly: AI could automate 25% of all work tasks in the U.S. and Europe, affecting an estimated 300 million jobs globally. These aren't fringe projections—they're consensus estimates from the world's most conservative institutions.
We broke all of this down in a recent episode of the What About AI? podcast. If you prefer to listen, here's the full conversation:
The World Economic Forum's Future of Jobs Report 2025 projected that 85 million jobs would be displaced by the shift to automation—but also that 97 million new roles would emerge. The Brookings Institution found that approximately 36% of U.S. jobs have “high exposure” to AI. The numbers were everywhere, but nobody was translating them into something an individual worker could actually use.
“When we started analyzing all 289 jobs, the patterns that emerged were not what most people would expect,” says James Perkins, co-founder of What About AI? and an enterprise technology leader with 18+ years in payments and fintech. “There were white-collar roles that a lot of people consider 'safe'—analyst positions, compliance jobs, certain finance functions—that actually scored higher on displacement risk than skilled trades. That was a wake-up call, even for me, and I spent years managing global payments teams.”
Each of the 289 professions in our database is assigned a displacement risk score ranging from 8% to 95%. That score isn't a guess. It's a weighted composite based on task repetitiveness, data dependency, physical-world requirements, interpersonal complexity, and the current state of AI tooling in that field.
“Most studies you see focus on theoretical full automation—could a machine ever do this job?” explains Sean Boyce, co-founder of What About AI? and a technology executive with over 15 years at Fortune 500 companies. “We weighted our factors differently because we were focused on near-term displacement—what's likely to happen in the next five to ten years with technology that either exists today or is on a clear development trajectory. That distinction matters enormously. It's the difference between a thought experiment and something you can actually plan around.”
After scoring all 289 professions, four patterns emerged consistently across every industry. These aren't theoretical frameworks—they're empirical observations from the data itself.
Task repetitiveness is the single strongest predictor of AI displacement. Jobs built around processing the same types of inputs and producing the same types of outputs—day after day, quarter after quarter—are exactly what large language models and automation agents are designed to handle.
Data entry clerks, for example, score 92% displacement risk in our database. The Bureau of Labor Statistics projects data entry clerk employment will decline by 35% by 2032—and that projection was made before the current wave of AI agents hit the market. Bookkeeping clerks (89%), payroll processors (85%), and basic accounting roles all follow the same pattern: high volume, low variance, structured data. These are the jobs AI handles first because they require the least adaptation.
If your entire job happens on a computer screen—if every input you receive and every output you produce is digital—your role is significantly more exposed. AI agents don't need robotic arms or mobility to replace screen-based work. They just need API access.
Telemarketers score 88% in our database. Customer service representatives in text-based channels score 82%. Basic report writers, scheduling coordinators, and first-tier technical support all cluster in the 75-85% range. The common thread: these roles exist entirely within the digital environment where AI operates natively.
Contrast that with an electrician (18%) or a plumber (15%). These jobs require navigating unpredictable physical environments, making judgment calls about aging infrastructure, and adapting to unique site conditions—capabilities that remain years away from reliable automation.
Roles that depend on empathy, trust-building, negotiation, or navigating complex emotional dynamics score consistently lower on displacement risk. A therapist (12%), a labor mediator (19%), a crisis negotiator (14%)—these professions involve reading context that AI cannot reliably interpret: body language subtleties, cultural nuance, the kind of emotional intelligence that comes from shared human experience.
This doesn't mean AI won't touch these fields. AI is already assisting therapists with session notes, helping mediators research precedents, and giving negotiators better data. But the core function—the human connection itself—remains irreplaceable in the near term.
This pattern catches people off guard. Creative professionals often assume they're safe because “AI can't be creative.” Our data tells a more nuanced story. AI is already generating competent copy, producing serviceable graphic design, composing background music, and writing functional code. The execution layer of creative work is increasingly automatable.
What remains protected is creative strategy—the ability to understand an audience, set a vision, make judgment calls about brand positioning, and connect creative output to business objectives. A creative director (28%) scores significantly lower than a production designer (67%). A marketing strategist (31%) is far safer than a copywriter focused on SEO content (74%).
The point of building this database was never to generate fear. Research from MIT Sloan and IBM shows that companies adopting AI are three times more likely to reallocate workers rather than eliminate them outright. The jobs aren't vanishing overnight—they're transforming. But transformation still requires you to adapt, and you can't adapt to a threat you don't understand.
“The point of all this work isn't to scare people,” says James. “It's that you cannot protect yourself from a risk you haven't taken the time to understand. When someone looks up their job in our database and sees a 72% displacement score, that's not a death sentence—it's a starting point. Now you know which parts of your role are at risk, and you can start building skills in the parts that aren't.”
Sean puts it in terms of his experience advising Fortune 500 companies: “I've spent 15 years watching technology transform enterprises from the inside. The people who thrive through these transitions are never the ones who ignored the data—they're the ones who confronted it early and repositioned. Whether you're an individual contributor or leading a team of 500, the playbook is the same: understand your exposure, then move toward the parts of your work that AI amplifies rather than replaces.”
The World Economic Forum's projection of 97 million new roles isn't just optimism—we're already seeing it. AI prompt engineers, model fine-tuning specialists, AI ethics officers, human-AI workflow designers, and AI implementation consultants barely existed three years ago. Today they're among the fastest-growing job categories.
Our database tracks these emerging roles too. The displacement scores for new AI-adjacent positions cluster between 8% and 22%—because they're built around capabilities AI needs humans for: judgment, oversight, cross-domain integration, and stakeholder communication.
The transition isn't painless. There's a real skills gap between the jobs being displaced and the jobs being created. That gap is exactly what our quiz, coaching program, and newsletter are designed to help you close.
If you take one thing from this analysis, let it be this: displacement risk is not destiny. It's a measure of how much of your current role overlaps with what AI does well today. A high score means you need to act—shift your focus toward the strategic, interpersonal, and creative dimensions of your work. A low score means you have time, but not infinite time.
Three concrete steps you can take right now:
| Claim | Source |
|---|---|
| ~30% of hours worked in U.S. could be automated by 2030 | McKinsey Global Institute, 2024 |
| 85 million jobs displaced, 97 million new ones created | World Economic Forum, Future of Jobs Report 2025 |
| AI could automate 25% of work tasks, affecting 300 million jobs | Goldman Sachs, 2024 |
| ~36% of U.S. jobs have “high exposure” to AI | Brookings Institution |
| Companies adopting AI are 3x more likely to reallocate workers | MIT Sloan / IBM Institute for Business Value |
| Data entry clerk employment projected to decline 35% by 2032 | U.S. Bureau of Labor Statistics |
| 289 jobs scored across 26 industries with displacement risk % | What About AI? Job Database |
Our database covers 289 professions across 26 industries. Every job has a displacement score, timeline estimate, and specific guidance on what to do next.
Coaching: For personalized 1-on-1 career guidance, visit whataboutai.com/coaching.
Take our free quiz to get a personalized assessment of how AI might impact your specific job and industry.
Take the Free QuizIBM announced it’s tripling entry-level hiring in 2026, targeting Gen Z workers. But they’ve rewritten every role for AI fluency, junior devs spend less time coding, and senior employees were let go. Is this a pipeline or a displacement?
Spotify’s co-CEO told analysts their top engineers have written zero code by hand since December. Internal platform HONK, built on Claude Code, lets engineers ship from their phones. 30% productivity gains, 50+ features, record margins. Here’s what it means for every knowledge worker.
Suleyman says 12–18 months. Amodei says 1–3 years. Altman, Musk, and Hassabis converge on the same window. These aren’t fringe voices — they’re the people building the technology. Here’s what it means for every knowledge worker.