After the Gaokao, How to Choose a Major Without Falling into AI's Traps?

If you've just walked out of the Gaokao exam room, congratulations. The next thing that will cause your whole family more anxiety than whether you finished your essay is: choosing your college major. AI has already shattered the logic of your parents' generation—"pick a stable, good major." But "studying AI guarantees a future" isn't true either. This article isn't going to list the "top 10 safest majors"—that kind of list changes every year. What it does is translate the most rigorous findings from 4 academic papers (the AIOE framework) about "which jobs are changing" into a major-selection decision you can actually discuss with your family.

This article builds on the site's existing Full AI Occupational Exposure Ranking: 5-Source Data Comparison dashboard (889 O*NET occupations × 5 independent AI exposure studies). That piece is the data foundation; this one is the application layer. They're best read together.

TL;DR · 30-Second Read

Core thesis: In the AI era, don't bet your major choice on "what AI can't do"—bet on "what lets AI work for you."

• What is AIOE: Four independent teams (Eloundou/OpenAI, Felten/Princeton, Webb/Stanford, Acemoglu/MIT) used four completely different methods to measure "how many tasks in each occupation will be affected by AI." Their conclusions are highly consistent: the most exposed are highly educated white-collar workers, not manual laborers.
• High AIOE ≠ Low income: Lawyers / financial analysts / accountants' BLS 2024 median annual salary remains in the U.S. top 25%; exposure measures "how much AI can help you," not "how much AI will put you out of work."
• The real moat: Licensing barriers (physicians/lawyers/teachers/registered engineers) + physicality + complex judgment = the triple combination that's hard to replace in the short term.
• Chinese context differences: U.S. entry-level white-collar jobs are collapsing, but China's AI industry is still expanding rapidly—in 2025, new AI job postings surged +543% YoY. Those enrolling in 2026-2030 will catch this wave, but the "hourglass effect"—where the top and bottom are stable while the middle gets hollowed out—is also playing out domestically.
• Counter-consensus insight: Graduates from top universities (Tsinghua/Peking/Fudan/SJTU) reading the "most unsafe majors" still have better employment outcomes than those from lower-tier universities reading the "safest majors." AIOE research completely ignores institutional signaling—this is its biggest blind spot.

§ 01 What is AIOE — 4 Papers, One Picture

AIOE = AI Occupational Exposure. It's not a single index but a set of methodologies: various ways to quantify the extent to which AI can replace or accelerate tasks within a job. Four teams, four independent routes:

1. Eloundou et al., GPTs are GPTs (OpenAI/Penn, 2023)

This is currently the most-cited "LLM × U.S. labor market" exposure study. The authors constructed a three-tier scoring system: E0 = LLM can't help; E1 = using ChatGPT directly can cut task time by ≥50%; E2 = requires specialized software layered on top of the LLM (code IDE, RAG, Agent) to cut ≥50%; the subsequently expanded E3 = requires image/multimodal capabilities.

Core finding: approximately 80% of the U.S. workforce will have ≥10% of their tasks affected by LLMs, and about 19% will have ≥50% of tasks affected; using LLMs alone can accelerate 15% of tasks, and with LLM-powered software this rises to 47-56%.

What this means for your major selection: Exposure measures "how much AI can help you," not "how much AI will make you redundant." A high E1 score for a role can be read as "mastering AI multiplies your efficiency."

2. Felten / Raj / Seamans AIOE Series (Princeton/NYU, 2018-2023)

Original methodology 2018 → formally published in Strategic Management Journal in 2021. They mapped 10 AI capability subfields tracked by the Electronic Frontier Foundation to 52 human abilities in O*NET (oral comprehension, inductive reasoning, hand-eye coordination...) via crowdsourced association, then projected onto 873 occupations.

The 2023 update for ChatGPT (Occupational Heterogeneity in Exposure to Generative AI) yielded a counter-intuitive conclusion: the most exposed are not factory workers, but literature professors. At the top of the exposure list are telemarketers, closely followed by post-secondary teachers in English literature, foreign literature, and history.

3. Michael Webb, The Impact of AI on the Labor Market (Stanford, 2020)

The third route: using AI patent text × O*NET task descriptions for semantic matching, counting how many tasks in each occupation overlap with existing AI patents. The most counter-consensus finding: unlike software/robotics, which primarily target mid-to-low skills, AI specifically targets high-skill tasks; if the trend continues, AI would actually narrow the 90:10 wage gap, but have almost no impact on the top 1%.

4. Acemoglu, The Simple Macroeconomics of AI (NBER w32487, 2024) — The Conservative

MIT's Acemoglu uses Hulten's theorem to calculate: AI's cumulative contribution to total factor productivity (TFP) over the next 10 years has an upper bound of only 0.53%–0.66%, annualized even less than 0.07%—far below Goldman Sachs's optimistic estimate of "GDP +7%, 300 million jobs replaced". His critique: early productivity experiments all focused on "easy-to-learn" tasks (customer service templates, code completion), but the truly difficult ones are hard-to-learn, context-dependent tasks, where there's no objective outcome measure and AI can't help much.

For parents guiding major selection: Even if you believe AI will reshape work, the total output shock over the next 10 years isn't as big as imagined—no need to panic-reshuffle your family's entire career plan.

A Comparison Chart

The four teams simultaneously have disagreements (Acemoglu is an order of magnitude more conservative than Eloundou) and consensus (highly educated white-collar workers are the most exposed). Anyone who tells you "AI will definitely replace XX major" or "AI will definitely not replace XX major" hasn't read any of these four papers.

A 2026 data point: The Anthropic Economic Index January 2026 report shows that 49% of occupations are already using Claude to complete ≥25% of tasks (up from 36% in January 2025); in November, the augmentation vs. automation ratio flipped to 52% vs. 45% (augmentation-dominant, not replacement). Actual usage data is currently falsifying the "AI replacement narrative"—at least for now.

Insertion point: After § 01 comparison table ========================================================= -->

§ 02 Translating AIOE Scores to the Major Level — High/Medium/Low Exposure Major Table

High-Exposure Majors (White-Collar "Knowledge" Work)

Brookings' Mark Muro, synthesizing Eloundou/Felten data in 2024, clearly states: high-paying, highly educated white-collar jobs are the most exposed—specifically non-programming roles in STEM, business and finance, architectural engineering, legal services, and mid-salaried office administration. The article specifically names bookkeepers, legal secretaries, HR assistants, bank tellers, and payroll clerks.

Pew Research 2023 re-scored using 41 O*NET work activities: 19% of U.S. workers are "highly exposed," 23% are "low exposure"; high-exposure examples: budget analysts, data entry, mechanical drafting, legal assistants, web development. Among college-educated groups, 27% are in the high-exposure tier, while among high-school-educated groups, only 12%—more education no longer automatically = safer.

Brookings also has a gender data point: 36% of women vs. 25% of men are in roles where "50% of tasks could be saved by AI"—because women are more concentrated in white-collar clerical positions.

Corresponding Chinese undergraduate majors (inferred from the AIOE framework): Accounting, Taxation, Auditing, Financial Analysis (non-trader track), Journalism, Advertising/Marketing, Translation, Law (basic compliance + document roles), Public Administration, Chinese Language and Literature (except teacher track), template-driven portions of Visual Communication Design.

Medium-Exposure Majors

Software Engineering / Computer Science / Data Science is a special case. Eloundou gives programmers the highest E1 score, but Brookings also notes this group is "AI-complementary" rather than "AI-replaceable"—because they write the AI tools themselves.

The actual data is even more counter-consensus: since ChatGPT launched, the total number of U.S. software developers has grown 7%, but recent CS graduates have an unemployment rate of 6.1%, higher than English (4.9%) and performing arts (2.7%). The total pie is growing, but the entry point is narrowing—seniors get promoted, juniors shrink.

Insertion point: After § 02 "Medium-exposure majors - Software engineering is a special case" paragraph ========================================================= -->

This pattern isn't limited to CS—medicine, design, and education are all experiencing the same internal divergence: senior doctors' wages are rising, while junior residents face unprecedented pressure; senior designers with AI tools are worth more, while junior designers are being replaced; veteran teachers have accumulated reputation, while new graduates face parents questioning "might as well let the child learn from AI."

Low-Exposure Majors

Pew explicitly lists the least-exposed: barbers, childcare workers, nannies. Brookings: blue-collar, physical work is "almost unaffected".

Corresponding Chinese undergraduate majors: Nursing, Rehabilitation Therapy, Stomatology, Clinical Medicine (surgical / emergency / anesthesiology / ICU tracks), Psychology (clinical counseling), Social Work, Preschool Education, Electrical Engineering and Automation (on-site construction type), HVAC / Water & Electric / Marine Electromechanical, Veterinary Medicine, Culinary Arts and Nutrition, Traditional Chinese Medicine Acupuncture and Tuina.

A Counter-Intuitive "High Exposure ≠ Unemployment" Example: Radiologists

Ten years ago, Geoffrey Hinton publicly predicted "we should stop training radiologists immediately"—but in 2026, the median salary for radiologists rose to $571K, and job numbers have grown 10% since ChatGPT; Jensen Huang used this example in a December 2025 public interview to rebut the panic that "AI destroys jobs".

The same story is playing out with paralegals: since ChatGPT launched, U.S. paralegal positions have grown 21%—despite Pew/Felten both rating it as high exposure.

This lesson is crucial for families making major selections: High AIOE means the "routine transactional work" in that profession will be taken over by AI, but the parts requiring professional judgment, trust, and accountability become even more valuable.

§ 03 Three Key Adjustments for the Chinese Context

All four papers above use U.S. O*NET data. Applying them directly to China will produce distortions, mainly due to three differences.

Adjustment 1: China's LLM Industry Is Still Expanding Rapidly

From January to October 2025, new AI job postings in China surged +543% YoY; algorithm engineer and LLM algorithm positions held steady at #1 and #2, while AI product manager roles grew 178% YoY. The average monthly salary for new AI positions is ¥61,764, 35.59% higher than the new economy sector average; LLM algorithm positions average ¥68,959/month, and AI scientists exceed ¥127,000/month; Maimai data shows over half of AI entry-level positions offer monthly salaries >¥50,000.

This is the biggest difference from the U.S. narrative of "entry-level white-collar job collapse." China's AI industry in 2024-2028 is in a "far more talent than needed" window. If you enroll in AI / Computer Science / Data Science / Automation / Electronic Information in 2026, you'll likely still be in this window when you graduate in 2030.

But a word of caution: this window is not permanent. Whether these positions will "continue expanding" or "begin saturating" after 2031, no one can guarantee you today. So your major choice must preserve transferability (see § 06).

Adjustment 2: Chinese Blue-Collar Wages Are Catching Up

In 2025, the average monthly salary for maternity matrons in China was ¥10,128; food delivery riders ¥8,325; truck drivers ¥8,279; food delivery riders' 3-year compound growth rate >10%. The blue-collar vs. white-collar average monthly income gap narrowed from a peak of ¥3,344 in 2013 to ¥2,250 in 2025 (−32.7%); blue-collar income growth has outpaced white-collar for 7 consecutive years.

In 2026, certified electricians/welders/forklift operators average ¥6,800-8,200/month, with core positions in tier-1 cities at ¥12,000+, 15%-20% higher than local ordinary administrative white-collar workers.

The U.S. mirror: After ChatGPT in 2022, the U.S. added 3 million white-collar jobs while blue-collar remained flat; but entry-level white-collar positions (requiring <1 year experience) dropped 50%. Both sides are playing out the same drama: "Top and bottom" are both stable, the "middle" is being hollowed out.

The implication for major selection isn't "make your child study to be an electrician"—but rather to acknowledge that the return on vocational education, associate degrees, and skilled trades is structurally rising, and "getting a bachelor's degree" need not be viewed as the only correct path. If family circumstances + the child's interests allow for high-quality vocational programs in mechatronics, auto repair, nursing, medical technology, or dental technology, the outcome might be better than forcing entry into a second-tier university's "safest major" bachelor's program.

Adjustment 3: Policy Moats — "Licensed Professions" Are Deeper and Less Transparent in China

Licensed physicians, lawyer's practice certificates, teacher certification, registered architects, registered geotechnical engineers, registered urban-rural planners, registered structural engineers, registered fire engineers, registered safety engineers, and more than a dozen others are all "access-type" professional qualifications—you cannot practice without the certificate, constituting a legal supply barrier.

These majors have legal protections in the Chinese context that cannot be directly replaced by AI. Unlike the U.S.: U.S. medical residency quotas are controlled by ACGME; China's are jointly controlled by the Ministry of Education + National Health Commission + Ministry of Human Resources and Social Security, making the moat equally deep but less transparent (meaning: higher barriers to entry, and the "certificate premium" after graduation is more stable).

The "Interim Measures for the Management of Generative AI Services" (jointly issued by seven ministries) taking effect in August 2023 introduced entirely new compliance positions at the AI regulatory level—algorithm filing, security assessment, data annotation, content review—these are among the few still-growing segments within law, cybersecurity, and media majors.

The Ministry of Education began promoting the "Four New" construction (New Engineering / New Medicine / New Liberal Arts / New Agriculture) in 2018, and released the "Reform Plan for the Adjustment and Optimization of Discipline and Major Settings in Regular Higher Education" in 2023—this is the state's policy alignment for "disciplines that lean on AI at the top and connect to the physical at the bottom." Among the newly added interdisciplinary majors, "Embodied Intelligence" has become the most watched new major, with 9 universities (including HIT and BUAA) opening it simultaneously.

§ 04 Five Counter-Consensus Insights — What You Think Is "Safe" Is Actually Most Dangerous

Counter-Consensus 1: High AIOE ≠ Low Income

May 2024 BLS OEWS data: U.S. lawyers' median annual salary $183,890; financial analysts $101,350; accountants and auditors $81,680—all well above the U.S. median of $49,500. These are all high-AIOE-exposure professions, yet their wages remain in the top 25% nationally.

So "high exposure" and "whether you should study it" cannot be directly equated. Felten himself repeatedly emphasizes in his paper: AIOE is exposure, not replacement; the same score is an opportunity for someone "willing to learn to use AI" and a threat for someone "passively waiting for the company to assign work".

Counter-Consensus 2: Complement vs. Replacement — Recent Graduates Actually Benefit Most

Brynjolfsson/Li/Raymond 2023 conducted a quasi-natural experiment at a Fortune 500 software company's call center: 5,179 customer service agents, after deploying a generative AI assistant, overall per-capita productivity +14%, novice and low-skill workers +34%, senior workers nearly zero.

This is the strongest evidence for "AI leveling the capability gap," and it's actually good news for recent graduates—AI makes entry-level workers capable of independent output in certain roles.

But the flip side of the same pattern is more alarming: companies are starting "experience creep"—positions requiring 2-4 years of experience dropped from 46% to 40%, while those requiring 5+ years rose from 37% to 42%—the result is "new hires can do the work, but companies no longer pay you to learn."

Insertion point: After § 04 counter-consensus 2 paragraph ========================================================= -->

The real advice for major selection: Choose a direction where you're willing to start using AI and producing output independently in your first year after graduation. The old path of waiting for companies to train you is narrowing.

Counter-Consensus 3: Major ≠ Occupation — AIOE Measures the Latter

A law graduate can become a lawyer, judge, compliance officer, HR professional, civil servant, corporate counsel, or government employee; a journalism graduate can become a reporter, PR specialist, corporate content creator, or product manager. AIOE gives "paralegal" a 0.9 high-exposure score, but the training law school provides (legal thinking, information structuring, contract negotiation) transfers to many occupations.

The real question when choosing a major should be: How many different occupations can the capability portfolio trained by this major transfer to? Not "what's the AIOE score of the most typical occupation corresponding to this major."

Counter-Consensus 4: "Betting on What Won't Be Replaced" Is the Wrong Question

Choosing IT in the 1990s seemed too niche; choosing finance in 2005 seemed safe; choosing civil engineering in 2015 seemed safe—all three decisions were proven wrong by the subsequent decade's reality.

Acemoglu himself cautions in his paper: early AI productivity data comes from easy-to-learn tasks; the impact on hard-to-learn tasks is unpredictable today. So rather than betting on a specific major being "definitely still around in 2035," it's better to choose directions with steep learning curves, transferable capability portfolios, and continuous certification update mechanisms.

Counter-Consensus 5: Top University Graduates in Any Major Outperform Lower-Tier University Graduates in the "Safest Major"

In 2025, the offer rate for Chinese master's/PhD graduates was 44.4%, undergraduates 45.4%—a degree inversion has appeared. But this is the full sample. Broken down by institution, undergraduates from top universities (Tsinghua/Peking/Fudan/SJTU) in "any major" have better employment outcomes than those from lower-tier universities in the "safest major."

Institutional signaling's weight in the Chinese labor market has not been systematically quantified by AIOE-type research—this is the biggest blind spot of the upstream AIOE ranking article, and a boundary this article must explicitly acknowledge.

The specific implication for major selection: Between "can go to a better school but need to change majors" and "can study the ideal major but at a weaker school," AIOE research doesn't tell you how to choose. But the reality of the Chinese labor market is: the former has a higher long-term return than the latter—especially when you can't be entirely certain what the world will look like in 2030.

§ 05 Decision Framework — Not Choosing "Irreplaceable" but Choosing "Complementary with AI"

Combining the four papers above + Chinese context + counter-consensus insights, three screening criteria for major selection:

Criterion 1: Can It Complement AI, Rather Than Be Replaced by AI?

Usable test questions: - Is the core capability of this major "organizing known information" (easily replaced by AI) or "making judgments under incomplete information + bearing responsibility" (AI can't replace)? - In the work you'll do after graduating from this major, does AI make you 1.5x more efficient, or does it eliminate your position?

Complementary type: Clinical medicine, surgery, clinical psychology, stomatology, rehabilitation, complex engineering (civil structures, power systems, nuclear engineering), teaching (especially K-12 + preschool), social work, nursing, emergency / firefighting, special education, archaeology, heritage restoration.

Risk type: Content moderation, basic accounting, entry-level translation, template-driven creative work, basic contract review, customer service copywriting.

Criterion 2: Is There a Structural Supply Barrier?

Licensing barriers + quota management + industry self-regulation together constitute labor supply rigidity. "Licensed majors" in the Chinese context: - Clinical Medicine (licensed physician) - Stomatology (licensed physician) - Law (lawyer's practice certificate) - Teacher Education (teacher certification + establishment quota) - Registered Architect, Registered Structural Engineer, Registered Geotechnical Engineer, Registered Urban-Rural Planner, Registered Electrical Engineer, Registered Fire Engineer, Registered Safety Engineer - Certified Public Accountant (though high AIOE, the CPA access barrier still constitutes a moderate moat) - First-Class Constructor, Supervision Engineer - Traditional Chinese Medicine (licensed TCM physician) - Veterinary Medicine (licensed veterinarian)

Beijing's International Professional Qualification Recognition Directory Version 1.0 is a window into "which certificates the government recognizes," which has reference value for major selection.

Criterion 3: Is There a "Hands-On Physicality + Complex Judgment" Combination?

Pew's least-exposed all share this characteristic: barbers, childcare workers, domestic helpers; Brookings explicitly states physical, non-routine blue-collar work has the lowest exposure.

Corresponding Chinese context: - Surgery / Emergency / Anesthesiology / ICU clinical medicine tracks - Stomatology (especially prosthodontics + orthodontics) - Clinical Psychology / Applied Psychology (counseling, correction, special education) - Rehabilitation Therapy / Sports Rehabilitation - Veterinary Medicine / Animal Medicine - Agriculture and Forestry (facility agriculture, smart agriculture, plant protection, horticulture) - Heritage Conservation / Archaeology / Restoration - Food Science (especially fermentation, flavor, sensory evaluation tracks)

§ 06 Specific Views on 9 Major Categories

Below, each category provides: Current AIOE risk assessment / 2026-2034 trend judgment / Complementary-type recommendations / Counter-examples. But please remind yourself: these are probabilistic judgments based on today's data, not prophecies.

Computer Science / AI / Electronic Information

• Current AIOE: Eloundou gives programmers the highest E1 score (easily accelerated 50%+ by LLMs), but they're also the core demographic that's AI-complementary
• 2026-2034 trend: In the Chinese context, AI job demand in 2025 surged +543% YoY; those enrolling in 2026-2028 will catch this wave; but after 2030, entry-level positions will undergo hourglass-ization
• Complementary-type recommendations: Artificial Intelligence, Computer Science and Technology, Data Science and Big Data, Electronic Information Engineering, Automation, Intelligent Science and Technology, Integrated Circuit Design, Embodied Intelligence (new interdisciplinary major in 2026), Cybersecurity
• Counter-example / Warning: Pure software engineering "application-layer CRUD" direction, traditional embedded rather than embodied intelligence direction—short-term employment is findable, but 5-10 year returns are lower than algorithm/system/chip directions
• Candidate universities (by tier): Tsinghua University · Peking University · University of Science and Technology of China · Shanghai Jiao Tong University · Zhejiang University · Nanjing University AI College · BUAA AI Research Institute · HIT School of Computer Science · Xi'an Jiaotong University AI College · University of Electronic Science and Technology of China

Medicine / Health

• Current AIOE: Diagnostic fields (imaging, pathology) have high exposure but jobs are actually increasing (radiologist salaries are rising); procedural fields (surgery, emergency, stomatology) have very low exposure
• 2026-2034 trend: China's aging population is a foregone conclusion, with structural growth in healthcare demand; the licensed physician certificate constitutes a deep moat
• Complementary-type recommendations: Clinical Medicine (surgical / emergency / anesthesiology / ICU tracks), Stomatology, Rehabilitation Therapy, Nursing, Anesthesiology, Medical Imaging (talent demand growing ≠ being replaced), Clinical Pharmacy
• Counter-example / Warning: Pure medical information management (high replacement risk), health services administration (going into administration rather than clinical); and "if you can study medicine, you should study medicine" is not unconditional—medicine's time cost is extremely high (bachelor's + master's + doctoral + residency training 11+ years), you need to be certain you're willing to bear it
• Candidate universities (by tier): Peking Union Medical College · Shanghai Jiao Tong University School of Medicine · Peking University Health Science Center · Fudan University Shanghai Medical College · Sichuan University West China Medical Center · Sun Yat-sen University School of Medicine · Zhejiang University School of Medicine · Huazhong University of Science and Technology Tongji Medical College · Central South University Xiangya School of Medicine · Capital Medical University

Education / Teacher Training

• Current AIOE: K-12 teachers have low exposure (high emotional/disciplinary/companionship components); university professors are rated by Felten as having higher exposure (lecture content is more structured)
• 2026-2034 trend: China's declining birth rate is a medium-to-long-term headwind for teacher education, but in the short term (5 years) there's still demand for establishment-quota positions; preschool + special education are actually structurally expanding
• Complementary-type recommendations: Preschool Education, Special Education, Primary Education, Psychology (applied / counseling / clinical tracks), Social Work
• Counter-example / Warning: Pure single-subject secondary school teacher track (fine short-term, but faces long-term K-12 enrollment shrinkage + AI content generation competition); non-teacher-education pedagogy (employment elasticity is lower than direct teacher education)
• Candidate universities (by tier): Beijing Normal University · East China Normal University · Central China Normal University · Northeast Normal University · Shaanxi Normal University · South China Normal University · Nanjing Normal University · Southwest University · Hunan Normal University · Capital Normal University

Law / Political Science / Public Administration

• Current AIOE: Legal assistants / paralegals have extremely high exposure, lawyers themselves have medium exposure; but paralegal positions have actually increased 21% since ChatGPT
• 2026-2034 trend: The lawyer's practice certificate remains a deep moat; AI compliance / algorithm governance / data compliance are new growth areas
• Complementary-type recommendations: Law (international / intellectual property / corporate / cybersecurity law tracks), Public Administration (civil service track), International Politics, Social Work
• Counter-example / Warning: Pure legal secretary track, assistant-type legal positions (the precise portrait AIOE warns about); purely liberal-arts-style public administration
• Candidate universities (by tier): China University of Political Science and Law · Peking University Law School · Tsinghua University School of Law · Renmin University of China Law School · East China University of Political Science and Law · Wuhan University Law School · Southwest University of Political Science and Law · Fudan University Law School · Shanghai Jiao Tong University KoGuan School of Law · Sun Yat-sen University Law School

Business / Economics and Management

• Current AIOE: Accounting / Auditing / Financial Analysis / Marketing Planning all belong to the high-exposure white-collar workers named by Brookings
• 2026-2034 trend: Licensed tracks (CPA, CFA) have moats; but pure marketing / planning / financial clerk positions will continue to shrink
• Complementary-type recommendations: Financial Technology (FinTech), Quantitative Finance, Actuarial Science (especially the actuary licensing track), Statistics, Insurance (especially underwriting + claims + actuarial tracks), Applied Economics
• Counter-example / Warning: Pure Marketing, Advertising, E-commerce, Business Administration—these are among the worst employment-feedback majors in Chinese undergraduate education, and the AI era will make them worse; unless your family can provide you with entrepreneurial resources + capital
• Candidate universities (by tier): Peking University Guanghua School of Management · Tsinghua University School of Economics and Management · Shanghai Jiao Tong University Antai College of Economics and Management · Fudan University School of Management · Shanghai University of Finance and Economics · Central University of Finance and Economics · University of International Business and Economics · Southwestern University of Finance and Economics · CUFE China Institute for Actuarial Science · Nankai University School of Finance (Actuarial)

Engineering (Non-IT)

• Current AIOE: Traditional civil / architectural engineering has low exposure (hands-on + on-site judgment), but China's architectural enrollment plans are down −6.1% YoY, already reflecting industry decline
• 2026-2034 trend: New energy / smart manufacturing / integrated circuits / aerospace / marine are policy priorities; traditional civil / architectural / chemical engineering continue to decline
• Complementary-type recommendations: Electrical Engineering and Automation, New Energy Science and Engineering, Energy and Power Engineering, Mechanical Design Manufacturing and Automation (smart manufacturing track), Aerospace Engineering, Naval Architecture and Ocean Engineering
• Counter-example / Warning: Civil Engineering (construction industry decline + real estate long-cycle downturn), traditional chemical engineering (environmental policy + capacity compression), traditional architecture (design demand compressed by AIGC + industry shrinkage)
• Candidate universities (by tier): Harbin Institute of Technology · Xi'an Jiaotong University · Beihang University · Tongji University · Tianjin University · Huazhong University of Science and Technology · Dalian University of Technology · Southeast University · University of Chinese Academy of Sciences · North China Electric Power University (Power specialty)

Humanities / History / Philosophy / Languages

• Current AIOE: In Felten's data, these are among the highest-exposure fields (second only to telemarketing)
• 2026-2034 trend: The job market is already contracting; but "humanistic literacy + professional skills" hybrids are actually scarce
• Complementary-type recommendations: Linguistics (computational linguistics track), Archaeology (hands-on), Heritage and Museum Studies, Teaching Chinese to Speakers of Other Languages (TCSOL + policy-driven), Anthropology
• Counter-example / Warning: Studying a single foreign language as an undergraduate major (already one of the fastest AI-replacement fields); pure Chinese Language and Literature / Philosophy / History in non-teacher-education tracks (extremely low employment elasticity). Exception: if your family has no rigid requirement for you to "use your degree to enter a specific path," these majors pursued to the depth of a doctoral degree can lead to research / publishing paths
• Candidate universities (by tier): Peking University · Fudan University · Nanjing University · Wuhan University · Sun Yat-sen University · Zhejiang University · Renmin University of China · Northwest University School of Cultural Heritage and Archaeology (Archaeology) · Shanghai International Studies University (Computational Linguistics) · Beijing Language and Culture University (TCSOL)

Agriculture / Forestry / Food / Environment

• Current AIOE: Core positions in agronomy, aquaculture, and forestry are all low-exposure; R&D + sensory evaluation in food science is low-exposure
• 2026-2034 trend: Policy dividends (New Agriculture + Rural Revitalization + Food Safety) + AI applications (smart agriculture) dual benefits
• Complementary-type recommendations: Veterinary Medicine (veterinary license), Horticulture, Plant Protection, Forestry, Food Science and Engineering (fermentation / flavor / nutrition tracks), Facility Agriculture / Smart Agriculture
• Counter-example / Warning: Traditional economic animal / economic plant tracks face industry consolidation pressure; non-applied environmental engineering (if not in drainage / atmospheric / solid waste / noise / monitoring tracks)
• Candidate universities (by tier): China Agricultural University · Northwest A&F University · Nanjing Agricultural University · Huazhong Agricultural University · Zhejiang University College of Animal Sciences · Ocean University of China (Marine Food / Fisheries) · Jiangnan University (Food Fermentation #1) · Northeast Agricultural University · South China Agricultural University · Beijing Forestry University

Art / Design / Media

• Current AIOE: Pure template-driven design (logos, posters, e-commerce hero images, entry-level voiceover, entry-level editing) has already been hit hard by AIGC; performing arts, pure instrumental performance, and parts requiring live presence have low exposure
• 2026-2034 trend: Pew/Brookings both specifically note design positions are being cut; but simultaneously performing arts recent graduate unemployment is only 2.7%, actually one of the lowest among all majors
• Complementary-type recommendations: Architecture (design + on-site construction + registered architect license), Urban and Rural Planning, Landscape Architecture, Industrial Design (tied to manufacturing), Theatrical and Film Performance (hands-on), Music Performance (hands-on), Heritage Restoration
• Counter-example / Warning: Visual Communication Design, Animation, Digital Media Art—unless your goal is "to become the top 5%," median income will be significantly compressed by AI; Broadcasting and Hosting (median voice actors are being undercut by AI pricing)
• Candidate universities (by tier): Central Academy of Fine Arts · China Academy of Art · Central Academy of Drama · Beijing Film Academy · Communication University of China · Tsinghua University Academy of Arts and Design · Tongji University College of Architecture and Urban Planning (Architecture + Landscape) · Southeast University School of Architecture · Central Conservatory of Music (Music Performance) · Shanghai Theatre Academy

§ 07 Reflections and Boundaries — Under What Conditions Does This Article Hold

The above covers specific views on 9 major categories, but each point has its boundaries. Before you use this article to argue with your family, put the following vulnerabilities on the table first.

1. AIOE Is "Capability Ceiling," Not "Employment Reality"

All four papers measure "how many tasks AI can help with," not "how many jobs AI is already replacing." There could be a delay of over a decade between capability ceiling and actual penetration (see the radiologist case). What you infer from AIOE as "this major will shrink in the next 10 years" might just be "certain tasks in this major will be accelerated".

2. AIOE Doesn't Account for Degree Signaling

Graduates from top universities reading the "most unsafe majors" still have better employment outcomes than those from lower-tier universities reading the "safest majors." Felten / Eloundou / Webb all fail to include school prestige as a variable in their models. This is a blind spot that Brookings and other reviews explicitly acknowledge.

Practical implication for major selection: Between "dropping down a school tier to get a safe major" and "aiming for a better school but studying a major with AIOE risk," please prioritize the school tier—unless you're already very clear about your life path.

3. AIOE Doesn't Account for Interest / Fit

Interest's impact on learning curves, persistence, and breaking through ceilings is enormous. A student with zero passion for biology who's pressured by parents into clinical medicine may not even become a licensed physician 11 years later. The AIOE perspective won't tell you this.

4. China's Job Market Lacks BLS-Level Granular Real Data

The China-specific data cited in this article (blue-collar wages, AI positions, graduate employment rates) comes from Zhaopin, Maimai, Xinhua, and new economy news—they all have caliber differences and sample biases. The kind of official "monthly median salary for every occupation" data that the U.S. BLS OEWS provides doesn't have a public Chinese equivalent. Therefore, predictions involving specific Chinese majors in this article have lower confidence than the U.S. portions.

5. The Five-Year Time Window Is a "Guess," Not a "Prediction"

2026 enrollment → 2030 graduation → 2035 is the first key career milestone. That's a 9-year window. Nine years ago was 2017—GPT-1 hadn't been born yet. Using 2026 AIOE data to predict the 2035 job market really doesn't have high confidence. That's why this article repeatedly emphasizes "choose transferable capabilities, choose complementary types" rather than "choose a specific major that's definitely safe."

6. Policy Moats May Narrow

Licensing systems for licensed physicians, teacher establishment quotas, registered architects, etc., are not immutable. China's teacher establishment reforms ("county-managed school-employed," "filing system"), multi-site physician practice, and expansion of constructor practice scopes are all in motion. What looks like a deep moat today may become shallower in 10 years.

7. This Article Doesn't Consider Your Family's Finances or Your Personal Aptitude

Medicine's 11-year program is feasible for upper-middle-class families but may not be sustainable for working-class families; art majors have rigid talent requirements; STEM majors have rigid math requirements; extroversion / introversion, stress tolerance, and team vs. independent work preferences all decisively affect major fit. The AIOE perspective assumes all students can equally choose any major—this isn't true.

One-Sentence Summary

Don't turn "choosing a major" into "betting on one that won't be replaced in 10 years." Make it "choosing a capability portfolio that lets you keep relearning + switching directions + complementing AI ten years from now". Specifically:

1. Licensed + Physical + Complex Judgment majors (clinical medicine, stomatology, nursing, rehabilitation, emergency, teacher education, registered engineer categories) are the most stable foundation
2. AI / Data / Algorithms / Integrated Circuits / New Energy—majors with dual policy + market dividends are suitable for reaching, but prepare for possible saturation after 2031
3. Traditional humanities / Pure design / Pure marketing / Pure clerical—high warning, but the top minority's path remains unchanged
4. School prestige priority is higher than major choice—unless you're completely certain of your life path
5. Major + Minor + One Certifiable Skill is the lowest-risk undergraduate structure

Further Reading

On-site: - Full AI Occupational Exposure Ranking: 5-Source Data Comparison — the data foundation for this article, cross-referencing 889 O*NET occupations - The Bursting of the Blue-Collar Myth: When Physical Labor Is No Longer Safe — the flip side of the blue-collar premium: which blue-collar jobs are being automated - The Accelerated Endgame for Entry-Level White-Collar Jobs — in-depth look at the "hollowed-out middle" white-collar side - Beyond UBI: Reimagining Distribution Systems — if AI truly reshapes employment, how should distribution systems be adjusted

Academic: - Eloundou et al. GPTs are GPTs (arXiv:2303.10130, OpenAI/UPenn 2023) - Felten/Raj/Seamans Occupational Heterogeneity in Exposure to Generative AI (SSRN 4414065, 2023) - Webb The Impact of AI on the Labor Market (Stanford 2020) - Acemoglu The Simple Macroeconomics of AI (NBER w32487, 2024) - Brynjolfsson/Li/Raymond Generative AI at Work (NBER w31161, 2023) - WEF Future of Jobs Report 2025 - Anthropic Economic Index January 2026

Chinese Policy and Data: - "Interim Measures for the Management of Generative AI Services" (CAC 2023) - Ministry of Education "Reform Plan for the Adjustment and Optimization of Discipline and Major Settings in Regular Higher Education" (2023) - Brookings: Generative AI, the American Worker, and the Future of Work - Pew Research: Which U.S. Workers Are More Exposed to AI

Author: Xiaoping Feng, [email protected]