Enrollment in dedicated AI degree programs has grown 45% annually for the past five years. Meanwhile, computer science enrollments dropped 7.7% in fall 2025 alone. That divergence captures a broader anxiety among students, parents, and career changers: which credentials will still be worth something when machines can draft contracts, write code, and diagnose diseases?
Anthropic CEO Dario Amodei has estimated that 100% of the code written at his company is now AI-generated, and he believes AI could eliminate half of entry-level white-collar jobs within five years. Hassan Taher, an AI consultant and author based in Los Angeles, has written extensively about how industries are absorbing these changes — and who stands to benefit.
Healthcare: High Demand, High Resistance to Automation
The U.S. Bureau of Labor Statistics projects that healthcare and social assistance will add approximately 2.1 million jobs from 2022 to 2032, more than any other sector, accounting for 45% of all new positions. Nursing, physician assistant studies, and allied health professions consistently rank among the most automation-resistant career paths.
Nurse practitioners, for example, face an estimated 0% automation risk, with projected job growth of 46.3% by 2031 and a median annual salary of $126,260. Physician assistants are similarly insulated. AI can analyze bloodwork and flag anomalies, but it cannot build rapport with a nervous patient, guide someone through a grief process, or make the kind of judgment calls that arise during a surgical complication.
Hassan Taher has expressed particular optimism about AI’s role in healthcare, but as a complement to human professionals rather than a replacement. Through his firm, Taher AI Solutions — a framing that positions human clinicians as essential partners, not redundant middlemen.
Trades and Construction: Hands That Machines Can’t Replace
Construction managers earn a median salary of $104,900 and face an automation risk of just over 3%. Electricians, plumbers, HVAC technicians, and welders occupy a similar position. AI can optimize scheduling, predict maintenance needs, and model structural loads, but it cannot rewire a house, repair a burst pipe, or operate in the unpredictable physical environments where most construction work happens.
The skilled trades have received less attention in conversations about AI-proof careers, partly because they don’t fit neatly into the four-year degree framework that dominates educational planning. But their durability is hard to dispute. Physical dexterity, real-time problem-solving in uncontrolled environments, and the ability to handle unexpected complications are exactly the qualities that current AI systems handle poorly.
Cybersecurity: Defending Against AI Requires Humans
AI is making cyberattacks more sophisticated, which paradoxically makes human cybersecurity professionals more essential. Hassan Taher has written about AI-enabled cyber threats, and the dynamic he describes creates a feedback loop: as offensive AI capabilities grow, so does demand for professionals who can defend against them.
Information security analyst roles are among the fastest-growing positions tracked by the Bureau of Labor Statistics. Defending against a complex, adaptive cyberattack requires creative thinking, real-time judgment under uncertainty, and an adversarial mindset — none of which AI does well on its own. Degrees in cybersecurity, or computer science with a security concentration, provide strong preparation for a field where demand reliably outpaces supply.
Law, Policy, and Governance: Protected by Institutional Design
Law offers an unusual form of job protection. AI can perform many of the cognitive tasks lawyers handle — document review, contract drafting, case research — and in some benchmarks does them competently. However, the practice of law is tightly regulated. Only licensed attorneys can give legal counsel, appear in court, and make fiduciary representations. AI cannot be licensed.
One analysis framed it this way: “An AI could pass the bar, but that doesn’t allow it to give legal counsel. The only people who can change this in the US are Congress, or the Supreme Court, two groups known for their close proximity to lawyers”. Political science and public policy degrees benefit from similar institutional protections — the people writing the rules about AI automation are the same people whose jobs automation would threaten.
Combining Depth With Technical Literacy
The World Economic Forum’s 2025 report offered a clear recommendation for students at every level: start with a core domain of interest, then add technical fluency. “Stack a minor or certificate in data science, computer science, cybersecurity, or operations analytics,” the report advised. “These supporting disciplines turn any major into a launchpad for AI-literate roles”.
Georgetown University’s Center on Education and the Workforce has tracked return on investment by major for years. Their data confirms that 17 of the 20 most lucrative majors come from STEM fields, with business and healthcare close behind. Since 2009, degrees conferred in the humanities and liberal arts have declined 33%, while education degrees dropped 14%.
Hassan Taher studied computer science at the University of Texas at Dallas before building a career that spans writing, consulting, and public speaking on AI — Taher has argued that the workers who thrive alongside AI will be those who understand how the technology operates and can direct it toward productive ends.
What Makes a Degree “AI-Proof”
No degree is entirely immune to disruption. But careers that require physical presence, emotional intelligence, licensed authority, creative judgment, or adversarial thinking in unpredictable environments hold advantages that current AI cannot easily erode. MIT Sloan research published in March 2025 found that many tasks being added to the U.S. labor market carry higher levels of human-intensive capabilities than the tasks they’re replacing. Rather than shrinking opportunities, AI may be redirecting them toward work that requires distinctly human input.
The safest bet isn’t a single major. It’s the combination of domain expertise, technical literacy, and the adaptability to keep learning as the tools change underneath you.