How many jobs is AI replacing in 2026? What the data actually shows

The headline number is 16,000 US jobs per month.

That's Goldman Sachs's April 2026 estimate of how many American workers are losing their jobs to AI-driven automation monthly. It's not a projection, it's based on current displacement data, with Gen Z workers taking a disproportionate share of the impact.

If you've been following AI job coverage and felt like the numbers never quite added up, that's because they don't, not without understanding the mechanism. This article covers what the April 2026 data says, why unemployment figures look calm while displacement is happening, who is actually being affected, and what it means for anyone currently in or entering an AI-exposed field.

The 16,000 number and what it means

Goldman Sachs's April 2026 analysis synthesizes displacement data across US industries and occupations. The 16,000 monthly figure represents net job cuts directly attributable to AI automation: positions eliminated or left unfilled because AI tools are now handling the work.

To put it in scale: the US labor market has roughly 160 million workers. 16,000 monthly job cuts represents 0.01% of the workforce per month, or about 192,000 workers annually. That's a rounding error at the macro level, which is why aggregate unemployment statistics have remained stable even as specific occupational groups feel meaningful pressure.

The number also doesn't capture the full picture of displacement. It counts direct eliminations. It does not fully capture the hiring slowdown, where positions that would have been created simply aren't, because AI is handling what those hires would have done.

Why unemployment data looks fine

This is the central paradox of 2026's AI displacement story, and it has a specific explanation.

Anthropic's March 2026 paper by Massenkoff and McCrory tracked unemployment rates for the most AI-exposed occupations versus the least exposed from 2016 through 2025. Their finding: no systematic increase in unemployment in the most exposed occupations since ChatGPT launched in late 2022.

The statistical framework they used could detect a "Great Recession for white-collar workers" scenario, where unemployment in the top exposure quartile doubled from 3% to 6%, and would show it clearly. That hasn't appeared.

What's happening instead is displacement through two quieter mechanisms:

Fewer separations, fewer hires. Employers are not firing their existing financial analysts, customer service teams, or data processors en masse. They are using AI tools to increase output per worker, which means when someone leaves or retires, they don't always replace them. The headcount shrinks gradually through attrition rather than layoffs.

Entry-level hiring compression. The starting positions that build careers, the data analyst role, the junior programmer job, the customer service associate, are being replaced by AI coverage of those specific tasks. Employers don't need as many people to do the entry-level work when AI handles a substantial portion of it.

Both mechanisms keep the unemployment rate stable while meaningfully compressing the labor market for certain workers.

For the full breakdown of which occupations are most affected and why, our AI job exposure by occupation guide covers the complete Anthropic data across 800 US roles.

Gen Z bears the brunt

This is the sharpest signal in the current data, and it's worth understanding precisely.

Brynjolfsson, Chandar, and Chen (2025) found a 6-16% fall in employment among workers aged 22 to 25 in AI-exposed occupations. Anthropic's team ran a parallel analysis tracking how often young workers (22-25) started new jobs in high-exposure versus low-exposure fields. Since ChatGPT's release in late 2022, the job-finding rate for this age group entering exposed occupations has fallen by roughly half a percentage point per month, amounting to approximately a 14% decline from 2022 baseline levels. The same effect does not appear for workers over 25.

Goldman Sachs's April 2026 analysis confirms this: Gen Z workers are disproportionately represented in the 16,000 monthly displacement figure.

The mechanism is specific. Young workers entering AI-exposed fields tend to start with the tasks that AI now handles most effectively: processing customer inquiries, generating reports from structured data, writing code to well-defined specifications, drafting routine documents. If employers are using AI to cover those task loads, they have less reason to hire the 23-year-old to do them.

This doesn't show up as youth unemployment, many of these workers are staying in school, shifting to lower-exposure fields, or staying in existing positions. It shows up as fewer entry points into careers in finance, programming, data, marketing analysis, and customer operations.

For workers currently in university programs targeting these fields, this is worth factoring into expectations about entry-level competition in 2026 and 2027.

Which fields face the most displacement

Not all AI-exposed occupations are experiencing the same pattern. The displacement is concentrated in fields where:

The tasks are well-defined and separable from human judgment. Data entry, customer inquiry handling, document processing, and code generation for specified requirements are all highly automatable in ways that broad analytical work or client-facing judgment is not.

The volume of work is high. AI tools produce the clearest ROI in roles where the same type of task is repeated hundreds or thousands of times, customer service, sales outreach, medical record coding, financial report preparation.

The on-ramp structure depends on junior employees. Fields where junior workers "learn by doing" the automatable tasks are seeing the most pronounced entry-level hiring compression, even when senior positions remain stable.

Anthropic's data puts computer programmers at 74.5% observed AI task coverage, customer service representatives at 70.1%, and data entry keyers at 67.1%. These three roles account for a substantial share of the employment volume in high-exposure occupational categories.

Notably, fields with high theoretical AI exposure but complex judgment requirements, lawyers (89% theoretical, ~15% observed), architects and engineers (84.8% theoretical, ~12% observed), and management roles (91.3% theoretical, ~20% observed), are seeing much lower actual deployment despite their theoretical vulnerability.

What AI displacement means for salaries

The compensation picture is more complex than "AI is lowering wages."

Workers in the most AI-exposed occupations currently earn 47% more than the least exposed group, $32.69/hr median versus $22.23/hr. High observed AI exposure is correlated with higher pay, not lower, because AI is concentrating in knowledge-work roles that have historically commanded premiums.

But within those roles, the wage structure is bifurcating. Workers who have developed demonstrable AI skills, prompt engineering, AI workflow design, output verification and quality control, are seeing premiums of 40-60% compared to peers without those skills. Workers performing the specific routine tasks that AI is absorbing are facing compensation pressure as supply-demand dynamics shift.

The practical question is not "does my industry pay well?" but "which specific tasks in my role are being automated, and is my skill profile above or below market median accounting for that shift?" That calculation also changes depending on where you work: a $95K salary in Austin covers meaningfully more ground than the same number in San Francisco. Our 2026 salary vs cost of living breakdown covers what the numbers look like across 19 major US cities.

PayScope benchmarks your specific resume, role, seniority, skills, and location, against active job postings, not survey averages. For workers in AI-exposed fields who want to understand whether their specific skills are trending toward premium or commodity, this level of specificity matters more than job-title averages. For a full analysis of how AI exposure is affecting compensation trajectories, see our guide on AI job exposure and salary impact.

The actual scale of what's happening

It helps to hold both parts of the story at once.

What's happening: AI is automating specific task categories across knowledge-work occupations at a pace that is compressing entry-level hiring, reducing headcount through attrition, and beginning to show up in monthly displacement figures (Goldman Sachs: 16,000/month). The displacement is real, concentrated in specific roles and age groups, and accelerating.

What's not happening: Mass layoffs of existing workers in high-exposure occupations. Systematic unemployment increases in the most affected fields. A collapse of white-collar employment comparable to what happened to manufacturing in prior decades.

The gap between these two realities is where most of the public confusion lives. Headlines about specific layoffs or AI-driven cuts compete with headline unemployment data showing stability. Both are true simultaneously.

The more useful question for any individual worker is not "will AI cause mass layoffs?" but "how is AI specifically changing the demand for my specific skills in my specific field?" The answer varies widely by role, seniority level, industry, and which tasks within the role are most automatable.

What entry-level workers and job seekers should do

If you're entering an AI-exposed field or currently early in your career in one, there are a few specific things the data supports.

Know which tasks in your target role are being automated fastest. Customer inquiry handling, data entry, routine document drafting, and code generation for specified requirements are the highest-automation tasks. Roles structured around these tasks face the most compression. Roles where the core work is judgment, client relationships, error-catching, or architectural decision-making face much less.

Develop demonstrable AI skill, rather than general AI familiarity. The wage bifurcation favors workers who can show they produce better outputs using AI tools, rather than workers who have merely used AI tools. The ability to design workflows, verify AI outputs, and handle edge cases that AI gets wrong is a concrete, marketable skill.

Check your market rate before accepting any offer. Entry-level salaries in some AI-exposed fields have been suppressed by reduced hiring competition. Knowing what comparable roles actually pay in your city for your specific skill profile, from job postings, not salary surveys, matters more in a compressed market than in a tight one. Upload your resume to PayScope to see where your profile currently sits against active job postings.

Consider whether your target field has growing or contracting entry-level demand. Software developers, despite high theoretical AI exposure, are projected by BLS to grow approximately 11% through 2034. Customer service representatives are projected to contract by roughly 5%. These projections reflect independent analysis of where AI displacement is and isn't changing employment demand.

Frequently asked questions

How many jobs has AI taken so far? Goldman Sachs's April 2026 estimate puts current AI-driven job cuts at approximately 16,000 per month in the United States. On an annual basis, that's roughly 192,000 jobs. The Anthropic Economic Index data also shows a 14% decline in job-finding rates for workers aged 22-25 entering high-exposure occupations since ChatGPT's launch, which represents hiring compression beyond direct layoffs.

Will AI cause mass layoffs? The current data doesn't support that conclusion as a near-term outcome. Anthropic's research found no systematic increase in unemployment in the most AI-exposed occupations since late 2022. Goldman Sachs's displacement figure of 16,000/month is real but represents a small fraction of total US employment. The more prevalent mechanism is hiring compression, fewer new positions, rather than mass termination of existing workers.

Which jobs is AI already replacing? The roles with the highest observed AI task coverage in 2026 are computer programmers (74.5%), customer service representatives (70.1%), and data entry keyers (67.1%). These are the fields where AI tools are most actively handling work tasks in professional settings today, according to Anthropic's usage data across 800 occupations.

Is Gen Z most affected by AI job displacement? Yes, the current data shows the clearest displacement signal among workers aged 22-25 in high-exposure fields. A 14% drop in job-finding rates for this age group, with no equivalent effect for workers over 25, suggests that entry-level positions in AI-exposed fields are contracting as employers use AI to handle what those hires would have done.

What jobs are growing despite AI exposure? Software developers (projected +11% through 2034), electricians (+11%), and healthcare roles with high physical or judgment components are among the fields with positive BLS employment projections despite AI's expansion. Productivity gains from AI tools appear to be expanding total demand in some software development contexts rather than eliminating positions.

How do I know if my job is at risk? The best starting point is understanding your specific role's observed exposure score from Anthropic's data (available in this site's AI job exposure guide), then assessing which specific tasks in your role are highest-exposure and whether your skill profile skews toward those or toward the judgment and oversight work AI is handling less. PayScope benchmarks your specific skills against current job posting data, which gives you a market-rate view that accounts for which capabilities are in demand right now.

The 16,000 monthly figure will likely grow as AI capabilities expand and adoption deepens. The workers best positioned in that environment are those who understand exactly which of their skills are above, at, or below market median, and can use that data to navigate compensation and career decisions before the market moves for them.

Upload your resume to PayScope; it's free, takes under ten minutes, based on active job postings. Know where your specific profile stands before your next negotiation or career move.