AI Risk Is Moving Toward White-Collar Work
A report highlighted by quasa.io says that 9.3 million U.S. jobs could be at risk from artificial intelligence within the next five years. Its most provocative claim is geographical: the sharper effects may land in high-tech hubs rather than in the manufacturing towns typically associated with economic disruption.
That matters because the usual career advice is increasingly outdated. For years, students and career changers were told to avoid manual or repetitive work and pursue a degree that leads to an office job—especially in technology, marketing, finance, consulting, or administration. AI is challenging that assumption. Many entry-level and mid-level white-collar jobs involve processing information, producing standard documents, summarizing research, writing code fragments, generating reports, or answering predictable customer questions. These are exactly the tasks generative AI can perform rapidly and cheaply.
The important distinction is that a job being at risk does not mean every worker in that occupation will be laid off. It can mean fewer new hires, smaller teams, slower wage growth, greater output expectations, or the removal of junior career ladders. For people choosing a degree or planning a pivot, that distinction is still serious: a profession can remain visible while becoming much harder to enter and less financially rewarding.
Why High-Tech Hubs May Feel the Impact First
High-tech metro areas have concentrated large numbers of workers in software, digital marketing, design, business operations, legal services, finance, media, and corporate support roles. These sectors employ people whose daily output is often digital, document-based, and measurable. That makes it easier for employers to insert AI tools into workflows.
The vulnerable pattern: standardized digital output
The most exposed work is not necessarily low-skill work. It is work with a repeatable structure and a digital final product. Consider these examples:
- A junior marketing coordinator drafting product descriptions, social posts, and campaign summaries.
- An entry-level analyst cleaning spreadsheets, preparing recurring slides, and writing first-pass market research.
- A support representative handling routine account questions through chat or email.
- A junior developer creating boilerplate code, documentation, tests, and simple integrations.
- A paralegal or legal assistant organizing documents, reviewing standard clauses, and producing routine summaries.
AI may not replace the need for a manager, strategist, engineer, attorney, or specialist. But it can reduce the number of people needed to complete the first 60% to 80% of a workflow. That is especially damaging to early-career workers because those routine assignments have traditionally been how they learned the business and proved their value.
The career-ladder problem is bigger than a single layoff
The deeper risk is not that every professional job disappears. It is that employers hire fewer beginners. A company that once hired five junior analysts may hire two analysts who can use AI well, plus one senior employee who reviews the work. The result is a bottleneck: graduates may hold relevant degrees but struggle to get the first two years of practical experience that employers demand.
This is why some degrees can become dead-end degrees in practice even when the subject remains useful. The issue is not whether a major has intellectual value. The issue is whether its graduate has a credible route into paid work, a portfolio of evidence, and skills that are difficult to automate.
Degrees Most in Need of a Career Plan, Not Blind Avoidance
No degree is automatically worthless, and broad claims that “AI will replace everyone” are not useful. Still, students should be cautious when choosing programs that lead mainly to generalist desk roles without a technical, regulated, client-facing, or operational specialization.
Degrees that may require a stronger complementary plan include general business administration, communications, marketing, journalism, graphic design, information systems without hands-on technical depth, and broad liberal arts programs with no applied experience. These fields can still lead to good careers. But a diploma alone is less likely to be enough when AI can generate competent first drafts of writing, visuals, presentations, and analyses.
A better question than “Will AI eliminate this major?” is: What task will an employer pay me to own after AI produces the first draft?
If the answer is vague—such as “I’ll be creative,” “I’ll work in business,” or “I’ll use AI”—the career plan needs more detail. If the answer involves accountability, field judgment, relationship management, regulated decisions, physical systems, sales, or implementation, the path is usually more durable.
Better Career Alternatives: Build Around Work AI Cannot Simply Finish
The best alternatives are not necessarily jobs that never use AI. In fact, resilient workers will use AI extensively. The goal is to work in roles where a person must validate the result, make a judgment under real constraints, persuade a stakeholder, operate equipment, or accept responsibility for an outcome.
1. Skilled trades and infrastructure careers
Electricians, HVAC technicians, industrial maintenance workers, plumbers, welders, elevator mechanics, and renewable-energy technicians work on physical systems in unpredictable environments. AI can assist with diagnostics, documentation, inventory, and training, but it cannot remotely repair a building’s electrical panel or install a heat pump.
These careers often have lower education costs than a four-year degree, clear apprenticeship routes, and demand linked to housing, grid upgrades, aging infrastructure, and energy efficiency. They are not effortless jobs; many require physical stamina, licensing, and safety discipline. But they offer a more direct connection between training and employability.
Registered nurses, radiologic technologists, respiratory therapists, dental hygienists, physical therapist assistants, sonographers, and many clinical laboratory roles combine technical competence with regulation, patient safety, and human interaction. AI will change documentation and diagnostic support, but licensed professionals remain accountable for care and execution.
For career changers, community-college and accredited associate-degree pathways may provide a better return than an expensive second bachelor’s degree. Before enrolling, check local licensing requirements, clinical-placement availability, completion rates, and actual job postings in your region.
3. Cybersecurity, systems, and AI implementation
“Tech” is not one labor market. Routine coding and generic digital production may face pressure, but organizations still need people who can secure systems, integrate software with messy business processes, manage cloud infrastructure, test AI outputs, govern data, and respond when systems fail.
The stronger path is not simply earning a generic computer science credential. It is building demonstrable capability: networking fundamentals, Linux, cloud platforms, identity management, security operations, databases, automation, and a portfolio of deployed projects. Employers increasingly value proof that a candidate can operate and troubleshoot real systems.
4. Revenue-facing and relationship-heavy roles
Complex B2B sales, account management, recruiting for specialized roles, procurement, partnership development, and customer success require trust, negotiation, context, and follow-through. AI can prepare account research and draft follow-up messages, but it does not own the commercial relationship when a six-figure contract stalls or a client needs an exception.
These roles are not safe merely because they involve people. Low-complexity telemarketing and scripted support remain highly automatable. The more resilient positions involve a long sales cycle, technical knowledge, domain expertise, and responsibility for retention or revenue.
A Practical Plan for Students and Career Changers
Do not respond to AI anxiety by collecting random certificates. Use a disciplined career audit instead.
Step 1: Break your target job into tasks
Read 20 current job postings and list the recurring duties. Mark which tasks are repetitive digital production, which require domain judgment, which require human interaction, and which carry legal, safety, or financial accountability. If most of the role is routine output, identify a specialization that moves you closer to decision-making or implementation.
Step 2: Pair academic study with a scarce capability
A communications student might add analytics, CRM administration, video production, public relations crisis work, or industry knowledge in healthcare or energy. A business student might add accounting systems, supply-chain operations, SQL, sales experience, or compliance. A design student might add UX research, accessibility, front-end implementation, or product strategy.
The goal is not to become a generic “AI expert.” It is to become useful in a specific business function where AI is a tool, not your substitute.
Step 3: Get proof before graduation
Internships, apprenticeships, clinical hours, freelance projects, campus IT work, volunteer operations work, and a public portfolio are more valuable than a vague claim that you are “proficient in AI.” Show an employer what you built, improved, sold, repaired, analyzed, or delivered—and explain how you checked the quality of AI-assisted work.
Step 4: Avoid debt without a labor-market case
Before taking on substantial tuition debt, calculate the likely first-job wage, completion risk, licensing requirements, and local demand. A prestigious program may be worth it for a profession with a clear placement pipeline. It is far riskier for a broad degree with no internship plan and no identifiable entry role.
The Bottom Line
The warning about 9.3 million jobs at risk should not be read as a prediction that work is ending. It is a warning that the value of routine knowledge work is being repriced. High-tech cities may feel the impact early because they have concentrated so much work that AI can accelerate, standardize, or partially automate.
For workers, the answer is not to run away from technology. It is to stop building a career around tasks that software can produce without understanding the real-world consequences. Choose a path with credentials, practical responsibility, domain expertise, human trust, or ownership of physical and operational outcomes. Then learn to use AI better than the person competing for the same role.
FAQ
Will AI really eliminate 9.3 million U.S. jobs in five years?
The figure describes jobs reported as being at risk, not a guarantee that 9.3 million people will be laid off. Risk can show up as reduced hiring, redesigned roles, fewer entry-level openings, or lower demand for certain tasks. Treat the number as a reason to assess your career exposure, not as a precise forecast of your personal outcome.
Which jobs are safest from AI?
No occupation is fully immune, but roles involving licensed responsibility, physical work in changing environments, complex human relationships, leadership, hands-on care, field service, and high-stakes judgment are generally harder to automate end to end. The strongest position is often a worker who uses AI while remaining accountable for the final result.
Should I avoid a computer science or marketing degree?
Not automatically. Avoid enrolling with the assumption that the degree alone guarantees a job. Computer science students should build systems, security, cloud, data, or domain-specific skills. Marketing students should add measurement, CRM, revenue operations, customer research, or sector expertise. Internships and portfolios are increasingly essential in both fields.
What is the best career move if my current job includes repetitive office work?
Start by identifying the parts of your role that require judgment, stakeholder trust, quality control, compliance, or implementation. Volunteer for those responsibilities, learn the AI tools affecting your workflow, and pursue training that moves you closer to operations, client ownership, technical administration, or regulated decision-making.
Fuente: quasa.io — Tue, 21 Apr 2026 07:00:00 GMT
