In 2020, the World Economic Forum predicted that 85 million jobs would be displaced by AI and automation by 2025. That number got plastered across every tech blog and career advice column for half a decade. Now that we're past that deadline, the results are in — and they're more nuanced than either the doomsayers or the optimists expected. Jobs were displaced. New jobs were created. And the salary implications for workers caught in the middle are profound.
The updated WEF projections for 2025-2030 are even larger: 92 million jobs displaced, 170 million created, for a net positive of 78 million new roles. But "net positive" is cold comfort if your specific role is on the displacement side of that equation. Let's break down where the risk actually sits, which roles remain insulated, and what this means for your paycheck.
The WEF Prediction: 85 Million Jobs Displaced
The original 2020 Future of Jobs Report projected 85 million roles would be displaced globally by machines and algorithms by 2025, while 97 million new roles would emerge — a net gain of 12 million jobs. The updated 2025 edition revised these figures upward significantly:
| WEF Projection | Jobs Displaced | Jobs Created | Net Change | |----------------|---------------|--------------|------------| | 2020 Report (by 2025) | 85 million | 97 million | +12 million | | 2025 Report (by 2030) | 92 million | 170 million | +78 million |
The net figures look reassuring. But they mask a critical detail: the displaced jobs and the created jobs require fundamentally different skills. A displaced data entry clerk doesn't automatically become an AI prompt engineer. The transition costs — retraining, geographic relocation, career pivots — fall on individual workers, not on the aggregate statistics.
What's happening on the ground matches the predictions directionally but with added urgency. 37% of companies surveyed now expect AI to directly replace human roles by the end of 2026. That's not a five-year horizon anymore. That's this year.
And the Harvard Business Review surfaced a troubling finding: companies are laying off workers based on AI's perceived potential, not its demonstrated performance. Executives are making preemptive cuts based on what they think AI will be able to do, not what it can reliably do today. This means some roles are being eliminated ahead of schedule, before the technology has actually proven it can replace them.
Most At-Risk Roles: Which Jobs Face the Highest Automation Exposure
Not all jobs face equal risk. The automation potential varies dramatically by role, and the data now gives us enough resolution to be specific:
| Role Category | Automation Potential | Salary Impact | |---------------|---------------------|---------------| | Data entry & processing | 85-95% | Severe decline (-15 to -30%) | | Customer service (tier 1) | 75-85% | Significant decline (-10 to -20%) | | Junior coding / basic development | 60-70% | Contracting market, flat pay | | Bookkeeping & basic accounting | 65-75% | Declining demand, stagnant pay | | Content writing (commodity) | 70-80% | Rate compression (-25 to -40%) | | Translation (standard) | 70-80% | Sharp decline in rates | | Basic graphic design | 50-60% | Template work commoditized | | Paralegal research | 55-65% | Reduced headcount, stable pay for survivors |
Junior coding roles face 60-70% automation potential — that's the headline number that's reshaping tech hiring. Entry-level hiring at the top 15 tech firms has fallen 25% year-over-year, and the trend is accelerating. Boilerplate code generation, simple bug fixes, test writing, and documentation — tasks that traditionally defined the junior developer experience — are now handled faster and cheaper by AI assistants.
Anthropic CEO Dario Amodei stated that AI will write essentially all code by the end of 2026. While that's likely an overstatement for complex systems programming, it's already close to true for straightforward CRUD applications, landing pages, and standard API integrations.
Key takeaway: The roles at highest risk share a common trait: they involve processing structured information according to well-defined rules. The more routine and rule-based your work, the higher your exposure.
Roles AI Cannot Replace (and Why They Pay More)
The flip side of displacement is that certain categories of work become more valuable as AI handles the routine tasks. These roles share characteristics that current AI systems fundamentally struggle with:
1. Complex System Architecture and Design
AI can generate code for individual components. It cannot design a distributed system that balances latency, consistency, cost, team capability, and business requirements across a three-year horizon. Architects and staff-plus engineers are seeing salary growth of 5-8% annually while generalist developer salaries stagnate at 1-2%.
2. Strategic Decision-Making Under Ambiguity
Product managers, business strategists, and senior executives make decisions with incomplete information, conflicting stakeholder interests, and long feedback loops. AI excels at optimization within defined parameters; it fails at defining the parameters themselves. C-suite and VP-level compensation continues to grow at 6-10% annually.
3. High-Stakes Human Interaction
Therapists, surgeons, trial lawyers, senior sales executives, crisis negotiators — roles where empathy, judgment, and real-time human reading are core competencies. These roles have seen minimal automation impact and in many cases face labor shortages, which pushes salaries upward.
4. Physical Work in Unstructured Environments
Electricians, plumbers, HVAC technicians, construction managers. Robotics for structured factory environments is mature, but robots that can navigate a century-old building's unique plumbing are decades away. Skilled trades wages have grown 4-7% annually since 2023.
5. AI/ML Engineering Itself
The people building, training, evaluating, and deploying AI systems are in extreme demand. AI engineer job postings are up 140% compared to 2024. The median AI engineer salary in the US now exceeds $185,000, with top-tier compensation packages at major labs surpassing $400,000 in total comp.
The pattern is clear: roles that combine domain expertise + judgment + human interaction are AI-resistant. Roles that are routine + digital + rule-based are AI-vulnerable.
The Salary Paradox: AI-Exposed Sectors Have Higher Wage Growth
This is the most counterintuitive data point in the entire AI-and-jobs debate. Research from the Federal Reserve Bank of Dallas found that sectors with the highest AI exposure are experiencing wage growth of 16.7%, compared to the national average of 7.5%.
Sectors most exposed to AI are paying more, not less.
The mechanism behind this paradox:
- Productivity amplification — Workers who effectively use AI tools produce 2-4x the output of those who don't
- Winner-take-all dynamics — Companies that adopt AI first gain market share, creating more revenue to distribute as wages
- Skill scarcity — The supply of workers who can effectively operate in AI-augmented environments hasn't caught up with demand
- Complementarity effect — AI makes skilled human judgment more valuable by handling the routine work and surfacing the decisions that need human input
This is not without precedent. The introduction of spreadsheet software in the 1980s didn't reduce accountant salaries — it eliminated bookkeeping clerks while increasing demand for financial analysts who could leverage the new tools. The same bifurcation is playing out with AI, just faster and across more occupations simultaneously.
Key takeaway: AI exposure correlates with higher wages for workers who adapt. The risk isn't AI exposure itself — it's AI exposure without AI competency.
Check how this dynamic affects salaries in your city using our salary calculator.
Entry-Level Workers Bear the Brunt
Across every industry — not just tech — the pattern holds: junior and entry-level roles absorb the most disruption. The reasons are structural:
- Lower task complexity makes roles easier to automate
- Higher labor cost sensitivity at scale (companies employ many juniors)
- Training investment looks less attractive when AI reduces the productive lifespan of entry-level skills
- Screening is harder — with fewer roles, companies raise the bar rather than expand the funnel
The numbers in tech are particularly striking:
| Metric | 2024 | 2026 | Change | |--------|------|------|--------| | Entry-level tech job postings (top 15 firms) | ~48,000 | ~36,000 | -25% | | P1/P2 hiring rate (industry-wide) | Baseline | -73% from 2023 | Severe contraction | | CS grad starting salary | $76,200 | $81,535 | +7% | | AI tool usage among developers | 72% | 84% | +12 pts |
The paradox of rising starting salaries amid falling hiring volume makes sense when you realize that companies are hiring fewer juniors but demanding higher capability from those they do hire. The $81,535 average starting salary for CS grads in 2026 reflects a more selective hiring process — only the strongest candidates get offers, and those candidates command higher pay.
84% of developers now use AI tools according to the Stack Overflow 2025 Developer Survey, up from 72% the prior year. For new graduates, AI tool proficiency isn't a differentiator anymore — it's table stakes. The differentiator is the ability to use these tools to solve problems that the tools can't solve alone.
Country-by-Country: Where Automation Hits Hardest
AI-driven job displacement doesn't affect all economies equally. The impact depends on economic structure, labor market flexibility, and the share of routine cognitive work in the employment mix:
| Country/Region | Automation Risk Level | Key Factors | |----------------|----------------------|-------------| | United States | High exposure, high adaptation | Strong tech sector, flexible labor market, rapid AI adoption | | United Kingdom | High exposure, moderate adaptation | Large financial services sector, growing AI investment | | Germany | Moderate exposure | Strong manufacturing (already automated), less routine cognitive work | | Japan | Moderate-high exposure, slow adaptation | Aging workforce, cultural resistance to displacement | | India | Lower near-term, high long-term | Large IT services sector at risk, but lower current AI adoption | | UAE/Gulf States | Moderate exposure | Diversifying economies, heavy investment in AI infrastructure | | Nordic Countries | Lower exposure | Already highly automated, strong retraining systems | | Southeast Asia | Variable | BPO/outsourcing sectors at high risk, manufacturing less so |
The US and UK face the highest near-term impact because their economies have large concentrations of the cognitive routine work that AI automates first — financial analysis, legal research, software development, content creation, customer service. But they also have the most flexible labor markets and the fastest AI adoption rates, which means the transition to new roles happens faster.
For workers considering relocation to optimize their career positioning, the country matters enormously. Our relocation guides cover the practical economics of moving between major employment markets, including visa requirements, tax implications, and true cost-of-living differences.
You can also use our cost of living comparison tool to see how your salary translates across different cities and countries in real purchasing power terms.
Five Career Moves to Make If Your Role Is at Risk
If your current role falls in the "high risk" or "medium risk" categories above, the time to act is now — not when the layoff notice arrives. Here are five data-backed moves:
1. Audit Your Task Mix
Break your daily work into tasks. What percentage is routine and rule-based versus judgment-heavy and ambiguous? If more than 60% of your work could be described to an AI system in a clear prompt, your role has high automation exposure. The fix isn't to work harder at those tasks — it's to shift your mix toward the judgment-heavy work.
2. Learn to Build With AI, Not Just Use It
There's a massive salary difference between using AI tools (+12% premium) and building AI systems (+43% premium). If you're a developer, move beyond Copilot autocompletion into fine-tuning, evaluation pipelines, RAG architecture, and AI system design. If you're in another field, learn to build AI-powered workflows, not just consume AI outputs.
3. Stack Domain Expertise on Top of Technical Skills
An AI engineer who understands healthcare compliance is worth significantly more than a generic AI engineer. A data scientist who understands supply chain logistics commands a premium over one who doesn't. Domain expertise is the moat that pure technical ability can't replicate, because it requires years of accumulated context that AI models don't have.
4. Negotiate From Data, Not Anxiety
Fear-based career decisions — accepting lower pay, staying in a declining role because it feels safe, avoiding negotiation — consistently lead to worse outcomes. Use market data to understand your actual position. Our salary percentile tool lets you see exactly where your compensation falls relative to peers in your role and location. Knowledge is leverage.
5. Consider Geographic Arbitrage
The salary gap between tech hubs and secondary markets is narrowing for remote roles, but widening for on-site AI positions. If you have the flexibility, comparing cities for your specific role can reveal opportunities where your skills command a premium that outpaces the local cost of living.
A Note on Timing
The World Economic Forum's timeline gives us through 2030 for the bulk of the projected displacement. But 37% of companies expect AI replacement to happen by end of 2026 — that's this year. The gap between corporate intentions and worker preparation is a window of opportunity for those who move early and a trap for those who wait.
The Bottom Line
The 85 million displaced jobs prediction was directionally correct, and the updated figure of 92 million by 2030 is backed by accelerating corporate adoption. But the net picture — 78 million new jobs created on top of those displaced — means this is a transition, not an extinction event.
The critical variables for your individual outcome:
- Role type matters more than industry. A routine cognitive worker in a "safe" industry faces more risk than a creative problem-solver in a "disrupted" one.
- AI competency is now a salary multiplier. The 16.7% wage growth in AI-exposed sectors goes to workers who can leverage the technology, not to those who merely coexist with it.
- Entry-level disruption is real and accelerating. The 25% drop in entry-level hiring at top firms and 73% contraction in P1/P2 roles is not a cycle — it's a structural shift in how companies build teams.
- The new jobs require different skills. A net gain of 78 million roles means nothing if you're trained for the roles being eliminated and not the roles being created.
- Geography shapes your exposure. Local AI adoption rates, labor market flexibility, and industry concentration all affect how automation impacts your specific market.
The workers who will thrive through this transition share three traits: they treat AI as a tool to master rather than a threat to fear, they invest in the judgment and domain expertise that AI can't replicate, and they make career decisions based on data rather than headlines.
The 85 million number was designed to scare. The 170 million new roles number was designed to reassure. Neither captures the reality for any individual worker. Your outcome depends on what you do in the next 12-24 months — and the data is clear enough to act on right now.
For a comprehensive look at how AI is transforming salaries, skills premiums, and career paths across the tech industry, explore our AI & Future of Work Salary Guide.