What White-Collar Jobs Will AI Replace?
Customer service, data entry, first-draft writing and coding, and junior analyst work are the white-collar tasks AI is displacing right now, and the evidence for it is measurable. This page looks specifically at what's being replaced, task by task, using real layoff and hiring data rather than industry-wide predictions.
Named companies have already cut real headcount citing AI as the reason. That's different from the broader "will AI eventually replace white-collar work" debate, and this page sticks to what's documented as already happening.
What the data actually says
Total job cuts across the US economy reached about 1.17 million in 2025, the highest level since the pandemic year of 2020 (CBS News). Of those, Challenger, Gray & Christmas tracked roughly 55,000 layoffs where the employer specifically cited AI as a factor (CBS News). That's a real, rising, named cause, still a minority of total layoffs but large enough to matter.
The named cases are specific and checkable. Amazon cut 14,000 corporate jobs in October 2025, followed by another 16,000 in January 2026, about 9 percent of its corporate workforce in three months, with CEO Andy Jassy explicitly connecting generative AI to the reductions (Yahoo Finance; CNBC). Salesforce CEO Marc Benioff confirmed the company cut 4,000 customer support roles with AI's help, shrinking that team from about 9,000 to 5,000, stating the company no longer needs to backfill support roles because AI agents handle the work (CNBC). IBM CEO Arvind Krishna told the Wall Street Journal that AI chatbots took over the work of a few hundred human resources employees, while IBM increased hiring in software engineering, sales, and marketing over the same stretch (CNBC). These are specific companies naming specific functions in specific quarters.
Which white-collar tasks are exposed right now
Four task categories show up across nearly every documented case:
Customer service. BLS projects a 5 percent employment decline for customer service representatives through 2034, citing automated self-service systems, apps, and social media reducing the need for a person to handle simple requests (BLS). Salesforce's 4,000-role cut in the same function, described above, is the clearest named example of that projection playing out as an actual layoff rather than a forecast.
Data entry and clerical work. Analysis of BLS projections found clerical typists facing a 36.1 percent decline rate, the steepest percentage drop of any occupation category, with office assistant roles projected to shrink by about 177,800 positions by 2034 (Visual Capitalist). This is the most straightforwardly automatable white-collar task category: structured, repetitive, and producing an output (a record, a form, an entry) that doesn't require a person's physical presence.
First-draft writing and entry-level code. BLS's own handbook for computer programmers states plainly that "computer programming work continues to be automated, with many companies expected to leverage technologies, including artificial intelligence (AI), to automate repetitive programming tasks," projecting a 6 percent employment decline for that specific occupation, one of the few times BLS names AI directly as a cause in its published text (BLS via what-jobs-will-ai-replace research). Software developers overall are still projected to grow 15 percent through 2034 (BLS), which means the decline is concentrated specifically in the entry-level, repetitive-task slice of coding work, not the discipline as a whole.
Junior analyst and paralegal-style document work. Paralegal and legal assistant employment is projected essentially flat, 0.2 percent growth, through 2034, with BLS attributing the limited growth directly to AI (BLS). The same pattern shows up in litigation support broadly: document review, sorting discovery for relevance and privilege, is one of the most widely adopted legal AI use cases today (American Bar Association).
The clearest single piece of evidence tying all four categories together is a Stanford Digital Economy Lab study using ADP payroll data: workers aged 22 to 25 in the most AI-exposed occupations, including software engineering and customer service, saw a 13 percent relative employment decline after generative AI adoption spread, growing to about 16 percent by October 2025 with no reversal yet, while more experienced workers in the same occupations saw no comparable drop (Stanford Digital Economy Lab). Employers aren't broadly firing experienced white-collar staff. They're not hiring as many entry-level people to do the tasks AI now handles.
Which white-collar tasks are protected, and why
The task categories above share a structural trait: repetitive, produces a written or numeric output, doesn't require an accountable decision-maker's signature. The tasks holding up share the opposite traits. A role that carries legal liability, requires licensure, or needs a human to own the consequences of a judgment call is not shrinking the way clerical and entry-level coding work is. A CPA's audit opinion, an attorney's filing, a licensed engineer's stamped drawing, all require a specific accountable person, and no AI tool holds a license or a bar number to put behind that accountability.
Client-facing advisory work is protected for a related reason. Companies cutting customer support and data-entry headcount, per the layoffs above, are simultaneously increasing hiring in roles that require judgment and relationship-building, exactly IBM's pattern of adding software engineering, sales, and marketing staff while cutting HR-adjacent transactional work in the same period.
What is already happening
The corporate layoffs named above are the clearest evidence: three named companies, Amazon, Salesforce, and IBM, cutting specific white-collar functions in 2025 and 2026 and citing AI directly as a driver in their own public statements. Dario Amodei of Anthropic and Ford CEO Jim Farley have both publicly suggested AI could eventually displace roughly half of entry-level white-collar positions, a claim that reflects executive expectation rather than measured outcome and should be read as such (CNBC).
At the same time, the Yale Budget Lab's ongoing tracking found no broad, economy-wide labor market disruption in the first 33 months after ChatGPT's release, and no clear relationship yet between an occupation's AI exposure and its unemployment rate through August 2025 (Yale Budget Lab). Both things are true simultaneously: specific named companies are cutting specific white-collar functions right now, while the broader labor market hasn't shown a wide fracture yet.
What to do about it
If your role is mostly customer service ticket handling, data entry, first-draft writing or coding, or document sorting, that specific bundle of tasks is the one companies are already cutting, not a future risk. The realistic response is shifting your actual daily work toward judgment calls, client relationships, and the parts of your role that carry some form of accountability, rather than waiting to see if your title gets eliminated.
Track your own task mix honestly over a normal week rather than reasoning from your job title. If more than half your time goes to the exposed pattern described above, repetitive, written or numeric output, no accountable sign-off, that's the signal to actively request different responsibilities now, before a restructuring decision makes that choice for you. If you're early career in one of the four exposed categories, the Stanford data says this is happening at the hiring stage first: seek roles or assignments that put you in front of clients, decisions, or ambiguous problems rather than execution queues.
Keep the judgment layer yours
The tasks you can still do without leaning on AI are what make you hard to replace here. The free 5-Day AI Reset is a five-email course built around exactly that: Day 2 has you take one task back and do it unassisted. One small change per day, and it stays useful no matter which way white-collar jobs move.
Frequently asked questions
What white-collar jobs is AI replacing right now?
Customer service, data entry and clerical work, entry-level coding and first-draft writing, and junior analyst or paralegal-style document review. Named 2025-2026 layoffs at Amazon, Salesforce, and IBM cite AI directly in these functions (CNBC).
Has AI actually caused documented layoffs, or is this speculation?
Both exist. Challenger, Gray & Christmas tracked about 55,000 US layoffs in 2025 where employers cited AI as a factor (CBS News), alongside named cuts at Amazon, Salesforce, and IBM. Broader economy-wide disruption hasn't shown up in Yale Budget Lab's tracking through August 2025.
Are entry-level white-collar jobs disappearing faster than senior ones?
Yes, in AI-exposed fields specifically. A Stanford study using ADP payroll data found a 13 to 16 percent relative employment decline for workers 22 to 25 in AI-exposed roles, with no comparable decline for more experienced workers in the same occupations (Stanford Digital Economy Lab).
Is customer service disappearing because of AI?
BLS projects a 5 percent decline for customer service representatives through 2034, citing automated self-service tools (BLS). Salesforce has already cut 4,000 support roles citing AI directly, a real reduction rather than a forecast.
Which white-collar jobs are safest from this trend?
Roles requiring licensure, legal accountability, or client relationship judgment: CPAs signing audit opinions, licensed attorneys filing under a bar number, engineers stamping designs, and advisory or client-facing roles. See What Jobs Are Safe From AI? for more.