Comparing Majors in the AI Era: Physics vs. Biology

“My kid in college is planning to transfer from biology to physics. Will his skills in physics be redundant in four years in this AI era?” This is a real question from a Chinese parent on an education forum. It’s the kind of question I hear more and more — from parents, students, and professionals trying to make career decisions in a world where AI is rewriting the rules.

It’s also the kind of question that deserves a better answer than “yes” or “no.”

The Real Question

Switching majors is always a big decision. In the AI era, it became a strategic one. The question isn’t whether physics skills will be “redundant” — it’s what kind of career resilience each major builds.

AI systems can already write code, analyze data, draft reports, and pass professional exams. Physics, as a quantitative discipline, appears to sit right in AI’s path. If AI can solve equations, run simulations, and generate analyses, does a physics degree still protect a career?

This concern is understandable but incomplete. It confuses what AI can do with tools and what AI cannot do with judgment.

Two Types of Resilience

Biology and physics don’t compete on the same axis. They build fundamentally different kinds of protection against AI disruption.

Biology builds role-anchored resilience. The roles themselves are hard for AI to penetrate — lab work, fieldwork, patient care, community health. These are embodied, regulated, and human centered. Biology graduates can move laterally within healthcare, environmental science, and policy. Most career paths are moderately protected, with lower variance in outcomes.

Physics builds person-anchored resilience. Protection comes not from the role but from the person’s cognitive toolkit — deep systems thinking, mathematical derivation, and the ability to operate across domains. Physics graduates can move into AI research, quantum computing, finance, and engineering. But the variance is higher: brilliant paths exist alongside exposed ones like generic data science.

What AI Will and Won’t Automate

Within the next four years, AI will automate much of the task-level physics work: solving textbook problems, running standard simulations, routine data analysis, and basic modeling. These are pattern-matching tasks that AI already handles well.

What AI will not automate in that timeframe is first-principles reasoning — deriving novel mathematical frameworks, identifying when established models break down, and constructing new approaches for unprecedented problems. This is exactly what physics training develops.

Large language models and their successors remain, at their core, statistical pattern engines. They are extraordinary at interpolating within known patterns. They struggle with genuine extrapolation — the kind of thinking that creates new knowledge rather than recombining existing knowledge.

The Decision Framework

So, should this student transfer? The switch to physics makes strategic sense if the student has a strong mathematical identity and finds derivation energizing rather than burdensome, is comfortable with higher-variance outcomes, wants maximum intellectual flexibility over credential-based career tracks, and aims to direct AI systems rather than perform tasks that AI automates.

Staying in biology makes sense if the student prefers more predictable career paths, values hands-on and human-centered work, and wants to enter fields where physical presence and human judgment are the core value.

Neither choice is wrong. They represent two different strategies for navigating the same disruption.

The Observer’s Insight

The parent’s question contains a hidden assumption: that skills are static things that either survive or become obsolete. But skills are cognitive toolkits that interact dynamically with a changing landscape. The real question isn’t whether a skill will be automated — many will. The question is whether the training builds the capacity to continuously create new value above AI’s rising capability floor.

In the AI era, there were two strategies for career resilience. Strategy A: go where AI cannot reach — embodied, regulated, human-centered work. Strategy B: think in ways AI cannot replicate — first-principles reasoning, novel framework creation, cross-domain synthesis.

Biology leans toward Strategy A. Physics leans toward Strategy B. Both are valid. The choice depends on who your child is, not on which field AI disrupts first.

What career decisions are you rethinking in the AI era? Leave a comment or reply to this email — I’d love to hear your story.

— Min Wu, PhD, Associate Professor,
Zilber College of Public Health, University of Wisconsin-Milwaukee