AI PhD 过剩了? 可能哦 ChatGPT answer may not be updated

本帖于 2025-08-27 23:25:39 时间, 由普通用户 RomaVacation 编辑

Here’s what I found on how many AI-related PhD degrees are awarded annually in the U.S., focusing on the most recent data available:


What the Data Tells Us

1. CRA Taulbee Survey (via AI Index / Our World in Data)

  • The AI Index Report—based on the CRA Taulbee Survey—provides data up to 2022 for AI PhD graduates in North America. The statistics for 2023, 2024, and beyond have not yet been published. Our World in Data

2. Degree Specialization in AI (2022–23 Academic Year)

  • In the 2022–23 academic year, there were 21,241 total PhDs awarded in computer science, computer engineering, and information fields—a record high. Of these, 28% specialized in AI or machine learning, highlighting the growing share of doctoral focus in AI. Inside Higher Ed

  • Applying that percentage yields roughly approximately 5,950 AI/ML PhD graduates for that year (21,241 × 0.28 ≈ 5,950).

3. AI-Relevant PhDs in Mathematics & Computer Science (2022 Estimate)

  • Another source reports 4,854 AI-relevant PhDs awarded annually in mathematics and computer science programs across U.S. universities, based on 2022 data. inkl


Summary Table

Year / Academic Year Estimated AI-related PhD Graduates
2022 (per “AI-relevant” metric) 4,854 annually inkl
2022–23 (based on 28% of CS/CE/I PhDs) ~5,950 estimated Inside Higher Ed
2023–2024 / 2025 No published data yet—awaiting updated reports Our World in DataCRA

Final Takeaways

  • The most recent official data (through 2022) suggests anywhere between 4,800 to 6,000 AI-focused PhD graduates per year.

  • The estimated range for 2022–23 (around 6,000) likely reflects the upward trend in AI specialization.

  • Unfortunately, 2024 and 2025 figures are not yet publicly available—we’ll need to wait for upcoming releases of CRA Taulbee data or the next AI Index report, expected around mid-2025. Our World in DataCRA


Let me know if you’d like to explore trends over time or how the data compares internationally!

 

That’s a really good and nuanced question.
Whether AI PhDs are in oversupply depends on how you frame the labor market: industry demand, academic opportunities, or geographic dynamics.


1. Industry Demand → Shortage, Not Oversupply

  • Tech companies, finance, healthcare, and government are hungry for PhD-level AI talent.

  • In the U.S., there were ~497,000 AI-related job postings in 2024 H1, but only ~193,000 qualified professionals—a 2.5× shortfall (jobspikr).

  • Growth is lopsided: demand for AI researchers expected to rise ~23%, while supply of PhD-level experts only grows ~4% (thecuberesearch.com).
    Conclusion: In industry, there is an undersupply, not oversupply.


2. Academia → Bottleneck / Perceived Oversupply

  • Faculty openings are scarce compared to the number of AI PhD graduates.

    • Example: some CS/AI PhD programs receive 250 applicants for 5 faculty-track spots ([reddit data point]).

    • Many new PhDs can’t find tenure-track positions and instead move to industry.

  • Result: oversupply in academia relative to tenure-track jobs, but not relative to total demand.


3. Geography & Immigration

  • A large share of AI PhDs in the U.S. are international students.

    • If visa/green card policies push them to leave, the U.S. feels a shortage.

    • If they all stayed, the U.S. academic market would feel more crowded.
      Immigration policy strongly shapes whether the U.S. feels “shortage” vs. “oversupply.”


4. Overall Balance

  • Industry & government → severe shortage of PhD-level AI experts.

  • Academia → oversupply relative to limited professor/research positions.

  • Macro picture (U.S.)net shortage of AI PhDs, especially in cutting-edge areas (deep learning, foundation models, AI safety, robotics, etc.).


Short answer:

  • No oversupply overall. The U.S. produces ~5,000–6,000 AI-relevant PhDs per year, but industry and government demand outstrips that.

  • Yes, oversupply in academia. Too many PhDs chasing too few faculty jobs.


Would you like me to make a side-by-side chart showing PhD graduates vs. job demand in academia vs. industry so the difference is clearer?

 

https://bidenwhitehouse.archives.gov/cea/written-materials/2025/01/14/ai-talent-report/ 

1/2 AI related PhDs are non us citizens. 

 

所有跟帖: 

打球的多,球星永远是少数 -米汤- 给 米汤 发送悄悄话 米汤 的博客首页 (0 bytes) () 08/27/2025 postreply 23:18:54

嗯,分化很严重 -bayview- 给 bayview 发送悄悄话 (583 bytes) () 08/28/2025 postreply 00:43:50

很奇怪的是,这还用讨论吗?什么行业都是这样,就是前几年CS那么热的时候 -Numero- 给 Numero 发送悄悄话 Numero 的博客首页 (412 bytes) () 08/28/2025 postreply 05:29:42

请您先登陆,再发跟帖!