Job Displacement | By 2030
Tech Sector Displacement by 2030: Uneven by Experience Level
Weighted average across 9 sources. Observed so far: ~11.8% (8 measurements from Yale Budget Lab, Brookings, Dallas Fed, BLS). Projections range 18.3–18.3% (median ~18.3%).
The sector-wide average of +10.4% obscures a K-shaped labor market split. Software developers aged 22-25 experienced a ~20% employment decline from late 2022 to Sept 2025 (Brynjolfsson, Chandar, Chen 2025), while workers aged 35-49 grew +9% over the same period. The primary mechanism is not mass layoffs but evaporated entry points: firms froze junior hiring as senior developers absorbed routine tasks with AI coding tools. Corroborating signals include a 30% drop in tech internships (Handshake), 25% fewer entry-level tech postings (SignalFire), and CS graduate unemployment at 7.2% vs. 4.5% for all graduates (ACS 2024). Meanwhile, BLS projects overall software developer employment to grow 17.9% through 2033, but the gains are concentrating among experienced workers who can effectively leverage AI. Note: Software has among the highest demand elasticity of any sector. When coding costs fall, firms fund projects that were previously unviable, potentially creating net new engineering roles even as per-project headcount shrinks.
TL;DR
The sector-wide +1.3% average hides a K-shaped split: workers 22-25 face -20% employment decline while workers 35-49 see +9% growth. AI tools amplify experience, so senior workers benefit while entry-level hiring evaporates. The average is technically correct but practically misleading for any individual.
The sector-wide +1.3% average is real but misleading. When broken down by age, the tech labor market reveals a sharp K-shape: early-career workers face steep decline while experienced workers are growing.
CORROBORATING SIGNALS FOR EARLY-CAREER IMPACT
WHY THE AVERAGE IS MISLEADING
The +1.3% sector-wide figure is a weighted average across all age bands. It is technically correct but practically misleading: a 22-year-old entering tech faces a fundamentally different labor market than a 38-year-old already in it. Policy and career decisions based on the average will be wrong for most individuals.
CAUSAL IDENTIFICATION
Brynjolfsson, Chandar & Chen use firm fixed-effects and health aides as a control group. After controls, the 22–25 cohort in AI-exposed jobs shows a 16% relative decline. The raw 20% figure includes ~4pp from macro conditions (Fed rate hikes, post-pandemic correction). The AI-specific effect is still large: roughly 4x the macro contribution.
THE MECHANISM: EVAPORATED ENTRY POINTS
This is not a story about mass layoffs. Firms stopped creating junior roles because senior developers, armed with Copilot, Cursor, Claude, can now absorb routine tasks that previously went to juniors. The traditional career ladder (learn by doing grunt work, get promoted) is being compressed. Junior hiring is a lagging indicator of how firms value human learning vs. AI output.
POPULATION WEIGHTS
From BLS SOC 15-1252 (CPS Table 11b, 2024), Stack Overflow 2024 demographics, ACS PUMS via DataUSA. Median US software dev age ≈ 33–34. The 22–25 band is 12% of the workforce but absorbs the majority of displacement.
WHO BENEFITS
Experienced developers (35-49) are seeing the strongest growth. AI tools amplify domain knowledge and judgment, exactly what senior engineers have. The productivity literature (Copilot RCT: +55.8% speed, Microsoft/Accenture: +26% PRs) mostly measures experienced developer gains. Early evidence suggests juniors benefit less from AI tools because effective use requires knowing what good code looks like.
BAND-LEVEL DETAIL
Primary: Brynjolfsson, Chandar & Chen (2025) “Canaries in the Coal Mine?” Stanford DEL / ADP Research. Corroborating: Dallas Fed (2026), Goldman Sachs (2025), CompTIA (2025), Handshake, SignalFire, BLS OOH 2024, ACS via Preston Cooper, ESB/Rabobank (2026, Netherlands).
Blended estimate across 9 sources ranging 0–22%. Higher-tier evidence and more recent data are weighted more heavily. See the full methodology for details on weighting, source validity, and recency bias.
Observed Data & Projections
This prediction has two fundamentally different types of evidence: observed employment data (what has actually happened) and forward-looking projections (what researchers estimate will happen). They are shown separately below because they answer different questions.
Filter by evidence tiers
What has happened
Measured employment data from government statistics, large-scale surveys, and administrative records. This is ground truth: what has actually occurred in the labor market.
Each dot is a different measurement source. Click any dot to jump to its source below.
What researchers project
Forward-looking estimates from structural models, institutional surveys, and expert forecasts. All projections target by 2030, shown by the reference line. The wide range (18.3–18.3%) reflects different model assumptions about reinstatement effects, demand elasticity, and adoption speed, not just parameter uncertainty.
Each dot is a different projection source. The x-axis shows when the projection was published. Click any dot to jump to its source. Overlay bars show directional signals from related studies.
Task Visualizer
What parts of your job will be cheaper to do with AI?
See which of your tasks face cost pressure from AI first.
Full Economy Picture
AI and the US Economy
Automation impact by occupation and income tier.
Sources (70)
Stanford HAI: software devs aged 22-25 employment down ~20% since 2024
Employment among software developers aged 22–25 has plummeted nearly 20% since 2024, even as their older colleagues' headcount grows. The pattern repeats in other jobs with higher levels of AI exposure, like customer service. Meanwhile, firm surveys indicate executives expect this trend to accelerate, with planned headcount reductions outpacing recent cuts.
Anthropic: coding 94% theoretical exposure but only ~30% actual task adoption
I was somewhat surprised that the gap between sort of coding in general, which as we point out had something like 94% theoretical exposure, but then based on actual adoption, it was closer to 30% of the tasks across all the jobs in that pocket of the economy.
BLS Mar 2026: Information sector 2,772K (+3K MoM), flat
Information sector employment at 2,772K, +3K month-over-month.
Chowdhury: 60% of closed CRA PRs have low signal quality
Autonomous coding agents are generating code at an unprecedented scale, with OpenAI Codex alone creating over 400,000 pull requests (PRs) in two months. CRA-only PRs achieve a 45.20% merge rate, 23.17 percentage points lower than human-only PRs (68.37%), with significantly higher abandonment. 60.2% of closed CRA-only PRs fall into the 0-30% signal range.
Stanford DEL: 88% coding productivity gain led to dev team cut from 7 to 3
At a private equity (PE)-owned company, an 88% productivity gain in coding led to reducing the development team from seven to three.
FRI: Moderate scenario → most 5-day freelance SW engineering jobs automatable by 2030
In the moderate AI scenario, almost all freelance software-engineering jobs requiring 5 days of effort from an experienced human are automatable by 2030.
the steepest risks in Information (18%)... Information 18.3%
Anthropic: Coding tasks migrating to automated API (+14% since Aug)
Since August 2025, the share of tasks in this category [Computer and Mathematical] has increased by 14% in the API and decreased by 18% in Claude.ai.
Fed: Coder employment flat post-ChatGPT after 4.8% pre-trend
After controlling for industry-level shocks we find that coder employment growth has been 3 percent lower since the introduction of ChatGPT. Cumulating over the roughly 3 years since November 2022 and using 5.735 million coder jobs as the base value, the implication is that roughly 500,000 additional coder jobs would have existed in the absence of large-scale LLM use.
GS Research (Briggs): Tech employment share has fallen below long-term trend
You can see AI's impact in the tech sector, where the employment share as a proportion of the whole economy has gone below the long term trend.
Lodefalk et al.: Sweden young software dev (22-25) employment shows steepest AI-exposure decline
The youngest cohort (22–25) shows the steepest decline in software developer employment, consistent with the economy-wide canaries effect. Software developers (SSYK 2512) represent a setting where AI primarily augments experienced workers.
Anthropic: Programmers highest AI exposure (74.5%) but no unemployment impact detected
Computer Programmers are at the top, with 75% coverage, followed by Customer Service Representatives
Challenger: Tech layoffs up 51% YoY in Jan–Feb 2026; AI cited as primary driver
Technology announced 11,039 job cuts in February for a total of 33,330 in 2026. That is an increase of 51% from the 22,042 cuts in this sector announced in the same period last year.
Ramp CPO: 500+ features shipped in 2025 with only 25 PMs
In 2025, Ramp shipped 500+ features, reached $1B revenue, and did it all with 25 PMs.
Citadel/Indeed: SW engineer job postings up 11% YoY (2026)
In spite of the current displacement narrative – job postings for software engineers are rising rapidly, up 11% YoY.
Cooper/ACS: 7.2% unemployment for early-career CS grads vs 4.5% overall (2024)
According to the American Community Survey, 7.2 percent of early-career college graduates who majored in CS were unemployed in 2024—well above the 4.5 percent rate for all early-career graduates.
Wang et al.: AI development heavily concentrated on CS/Math — 7.6% of employment
software engineering abilities has the potential to accelerate other varieties of work. However, this concentrated focus significantly overrepresents a domain that accounts for only 7.6% of total employment.
MIT FutureTech: 55.9% AI task success in computer/math roles (N=1,955)
Computer and Mathematical job family: 55.9% AI task success rate across 1,955 evaluations, with statistically significant negative slope (beta=-0.25) indicating longer tasks are harder for AI. Based on worker evaluations of LLM outputs on real O*NET tasks.
Goldman (Mei): AI impact 'largely confined to tech sector'; expects it to grow materially
AI's labor market impact still small — only 5-10K/month drag on net job growth. Only 2½% of jobs exposed to automation today. Baseline forecast: 6-7% of workers displaced (range 3-14%), lowering annual hiring by 1M jobs and raising unemployment ~½pp. US creates 30M gross new jobs/year; does not anticipate job apocalypse. Impact so far largely confined to tech sector but expects it to grow materially.
Dallas Fed: Computer systems design employment down 5% since ChatGPT release
Employment in the computer systems design and related services sector has declined 5 percent.
METR (updated): AI made experienced open-source devs 18% slower (CI: -38% to +9%)
METR's follow-up study finds original 19% slowdown unreliable due to selection effects. New data shows -18% speedup (CI: -38% to +9%) for original developers and -4% speedup (CI: -15% to +9%) for new recruits. Study redesigned due to severe selection bias.
Yale Budget Lab: Computer/math occupations score 8.4/10 exposure (highest of 22 groups, 342 BLS occupations)
Computer and Mathematical occupations have the highest average PCA-weighted exposure score (4.2) and highest variance (0.739). GPT-scored exposure for SOC 15-0000: mean 8.4/10, range 8-9 across 10 occupations.
CIO/Robert Half: 61% of IT leaders plan to increase permanent headcount in H1 2026
37% of tech professionals say their role has been 'redefined or restructured due to gen AI in the past two years.' 61% of IT leaders say they plan to increase permanent headcount in the first half of 2026.
Deutsche Bank: 85% of devs use AI coding assistants; up to 60% productivity gains
Over 85% of software developers now use AI coding assistants, delivering productivity gains of up to 60%. This shift raises questions about traditional software engineering roles and licensing models.
Indeed: Tech postings 34% below pre-pandemic; AI tech postings 45% above
the number of tech postings that mentioned AI was about 45% higher than in February 2020, while total tech postings were 34% below pre-pandemic levels
ESB/Rabobank: Dutch ICT youth employment fell 11% despite 6.5% sector growth (since 2022)
De ICT groeide in dezelfde periode met 6,5 procent, maar zag een daling van de werkgelegenheid van zo'n 11 procent onder 15- tot 24-jarigen.
Anthropic: Computer/math tasks ~33% of Claude.ai conversations, ~50% of API traffic
Computer/mathematical tasks account for ~33% of all Claude.ai conversations and ~50% of API traffic, indicating concentrated AI impact on tech-adjacent roles.
Cognizant: computer/math roles 67% task exposure — second highest of all categories
Education task exposure jumped from 11% to 49% — a 4.5x increase driven by multimodal AI, advanced reasoning, and agentic systems. Based on reassessment of 18,000 tasks across 1,000 O*NET occupations.
IMF: High IT skill demand won't translate 1:1 to specialist jobs as AI automates IT tasks
The high demand for new IT skills may not necessarily translate into a one-for-one rise in demand for IT and AI specialists, especially as many IT tasks may progressively be automated by AI.
BLS: Computer programmer roles projected to decline further 6% (2024-2034)
BLS projects a further 6% decline in programmer roles through 2034. Computer programmer employment (routine coding roles) fell ~27.5% in roughly two years following ChatGPT's release — one of the largest two-year drops in any occupation tracked by BLS. Software developer employment remained flat.
Dallas Fed: young worker share fell 0.9pp
Employment share of young workers (20-24) in most AI-exposed occupations fell from 16.4% to 15.5% between November 2022 and September 2025.
Chen & Stratton: 8.5% coding boost, 8.7% faster tasks, no employment effects
AI adoption led to moderate productivity increases — an 8.5 percent increase in coding activity and 8.7 percent faster task completion — with no measurable quality declines. These productivity gains did not translate into increased output, changes in task composition, or effects on employment.
Challenger: 54,694 AI-cited job cuts in 2025
54,694 job cuts cited AI as a reason in 2025, concentrated in tech, finance, and customer service sectors.
MIT Iceberg Surface Index: 2.2% of tech wage value ($211B, 1.9M workers) exposed to AI
Nationally, the Surface Index stands at 2.2%, representing approximately $211 billion in wage value across 1.9 million workers in technology occupations.
Early-career workers (ages 22-25) in AI-exposed occupations experienced 16% relative employment declines, controlling for firm-level shocks. By September 2025, employment for software developers aged 22-25 declined nearly 20% compared to its peak in late 2022.
Sarkar (UChicago/Cursor): 39% more code merges after AI agent became default mode
Software output increased by 39% after the AI agent became the platform's default code generation mode. Quality did not deteriorate. Senior developers benefit most from agents; junior developers benefit most from autocomplete.
CAIS/Scale AI: AI fails 97.5% of real-world remote work projects at professional quality
Failures predominantly stem from file errors (18%), incompleteness (36%), sub-professional quality (46%), and inconsistencies (15%). Best agent (Manus) achieved only 2.5% automation rate across 240 real freelance projects.
Challenger: 78K tech jobs lost to AI in H1 2025
78,000 tech sector job losses attributed to AI in the first half of 2025, with QA testing, junior development, and IT support roles most affected.
BLS: SW dev growth revised down from 25% to 17%
Software developer employment projected to grow 17.9% (2023-2033). Despite AI exposure, this occupation is unlikely to experience employment decline because robust software needs are expected to support continued demand.
CompTIA: IT unemployment rose from 3.9% to 5.7%
IT unemployment rate rose from 3.9% to 5.7% between late 2024 and mid-2025, the highest level since 2021, driven by AI-related restructuring.
Anthropic: 37% of software dev tasks show AI involvement; writing code is 18% of usage
37% of software development tasks show AI involvement. Writing code is 18% of actual AI usage in software development; most use is for understanding, debugging, and reviewing code.
Goldman: Young tech workers (20-30) unemployment risen ~3pp since start of 2025
Young tech workers (20-30) unemployment risen ~3pp since start of 2025; tech employment share fallen below pre-pandemic trend.
St. Louis Fed: Computer/math occupations (80% AI exposure) show steepest unemployment rise
Occupations with higher AI exposure experienced larger unemployment rate increases between 2022 and 2025 (r=0.47). Computer/math occupations (exposure ~80%) saw steepest rises. GenAI-intensive occupations: r=0.57.
Overall tech postings 36% below pre-pandemic baseline as of July 2025. Software engineer postings down 49%. Annualized over ~2.7 years: ~13.5%/yr decline rate.
OECD: Employment Outlook 2025 on AI and the labour market
JPMorgan: Tech/info/media employment stopped growing at end of 2022
Tech/info/media employment stopped growing at end of 2022, coinciding with ChatGPT release; less than 10% of firms using AI regularly as of mid-2025.
Microsoft/Accenture: 26% more pull requests with Copilot
26% increase in completed pull requests, 13.55% increase in commits, 38.38% increase in builds — but build success rate fell by 5.53 percentage points.
Daniotti et al. (Science): 30% of US Python code AI-assisted; 3.6% output increase
AI-assisted coding jumped from around 5% in 2022 to nearly 30% in the last quarter of 2024. Quarterly output has increased by 3.6%. Experienced programmers benefit more; beginners hardly benefit at all.
BLS projects computer programmer employment to decline 9.6% from 2023-2033 (~1% per year). Software developers projected to grow 17.9% over the same period.
BLS OOH: Computer programmer employment projected to decline 9.6% (2023-2033)
Overall employment of computer programmers is projected to decline 9.6 percent from 2023 to 2033.
Software developer job postings 33% below pre-pandemic baseline as of November 2024. Annualized over ~2 years since ChatGPT launch: ~16.5%/yr decline rate.
Alphabet: AI tools reduced need for certain engineering and support roles
Alphabet disclosed ongoing workforce restructuring, noting AI tools have reduced need for certain engineering and support roles while increasing demand for AI/ML specialists.
Meta: 'Year of Efficiency' cut ~21K jobs; AI-driven productivity enables smaller teams
Meta's 'Year of Efficiency' resulted in ~21,000 job cuts across 2023. Company cited AI-driven productivity gains enabling smaller teams to maintain output.
HBS: 17% decrease in postings for top-quartile automation-potential occupations
17% decrease in job postings for occupations in top quartile of automation potential, alongside 22% increase in postings for augmentation-prone occupations.
Forrester: AI and automation will take 6% of US jobs by 2030
Job openings grew 80% for AI scientists and 70% for ML engineers, while backend engineer postings declined 14%, frontend 24%, and data engineers 20%+.
Computer programmer employment (routine coding roles) fell ~27.5% in roughly two years following ChatGPT's release (~13.8%/year annualized) — one of the largest two-year drops in any occupation tracked by BLS. Software developer employment remained flat.
Paradis et al. (Google): 21% reduction in coding time per task
They found a 21 percent reduction in time spent per task. Interestingly, more experienced developers saw bigger effects, possibly because AI requires substantial verification and judgment rather than simply accepting generated code.
HBS/Hoffmann: AI-generated code has 41% more vulnerabilities; devs overestimate security
AI-generated code has 41% more vulnerabilities than human-written code. Developers overestimate the security of AI-generated code.
Yeverechyahu et al.: GitHub Copilot → 37-55% increase in commits
They found a 37-55 percent increase in commits, mainly through contributions building on others' work like debugging or small edits.
Gambacorta et al.: CodeFuse → 55% increase in code output at Ant Group
They show that 20 percent of the 55 percent total increase was directly attributable to AI-generated lines, suggesting genuine productivity effects beyond simple code generation.
LinkedIn/Baird: Copilot users show increased code output but no displacement evidence
GitHub Copilot users show increased code output but no evidence of displacement in hiring patterns.
Acemoglu: ~5% of tasks fully automatable; job impact more modest than claimed
AI will increase productivity but impact on jobs more modest than often claimed; estimates ~5% of tasks fully automatable.
Current employment data do not yet show large-scale AI-driven displacement.
Software development postings 28% below pre-pandemic baseline (Feb 2024). Annualized over ~1.3 years since ChatGPT launch: ~22%/yr decline rate.
MIT: Only 23% of vision-task wages cost-effective to automate with AI today
Only 23% of worker wages for vision tasks would be cost-effective to automate with AI today.
Goldman Sachs: AI could expose 300M full-time jobs globally to automation
AI could expose 300 million full-time jobs globally to automation.
McKinsey: GenAI could automate up to 70% of business activities across occupations
Generative AI could enable automation of up to 70% of business activities across occupations.
WEF: 44% of workers' core skills will be disrupted in the next five years
Employers estimate that 44% of workers' core skills will be disrupted in the next five years.
GitHub Copilot RCT: developers 55.8% faster
RCT: developers completed a coding task 55.8% faster with GitHub Copilot assistance.
Know a study we’re missing? Suggest a source for this prediction.