ChinAI #110: So You Want to be an AI Major?

Plus, Breaking down Ant Financial's Prospectus

Greetings from a world where…

(one of the best Twitter follows)

…As always, the searchable archive of all past issues is here. Please please subscribe here to support ChinAI under a Guardian/Wikipedia-style tipping model (everyone gets the same content but those who can pay support access for all AND compensation for awesome ChinAI contributors).

Feature Translation #1: the “AI Major” in China

Thanks everyone for taking part in last week’s Around the Horn. Really close call among 4, 6, 7, and 8. I chose 8 (AI majors) since it had the most support among the choices that paying subscribers voted for, and 4 (Ant Group’s IPO) because it had the most support overall. Was helpful to see that almost every single option received at least some interest — except #9 about the early history of Hongqi cars (I’m glad I did the ATH exercise because there’s a good chance I would have gone for that one)

So…you want to be an AI major?

Context: High school seniors took the 2020 college entrance examination (gaokao) in July, and now they are thinking through which school and major they should apply for. One especially hot major is AI.

  • In March 2019, 35 colleges and universities across the country, including University of Science and Technology Beijing, Shanghai Jiaotong University, and Nanjing University, obtained the first batch of qualifications for establishing a new major of “Artificial Intelligence.”

  • Currently, a total of 215 universities across the country have added AI majors.

The choice of major matters a lot. One crucial aspect of China’s college admissions system: each university sets aside a number of places for each major, which means students apply to a specific major of a university. So, even if your scores are above the average successful applicant for a university, if you apply to an especially competitive major (which may be the case for new AI majors), you may not get in at all because your score was not high enough for the cutoff in that particular major.

Southern Metropolis Daily took a look at the curriculum, salaries, and employment prospects for AI majors. Key Takeaways:

  • Curriculum: 烧脑  ("brain-burning"), demands much thinking and reasoning, relatively large proportion of math classes. Here’s the math schedule for the AI major at my dad’s alma mater, Shanghai Jiaotong: Advanced Mathematics, Linear Algebra, Discrete Mathematics, Probability Statistics, Stochastic Processes, Linear Optimization and Convex Optimization. AI majors are oriented toward “elite cultivation”: about 30-40 people enrolled in AI majors from publicly available data at various schools.

  • Employment prospects/Salaries: Per a "2019 China AI & Big Data Talent Employment Trend Report" released by Liepin (big online talent services platform), the average monthly salary of AI and big data talents is 22,322 RMB. Data architects and data scientists rank the highest (37,451 RMB and 36,570 RMB). The average monthly salary of positions in deep learning, image recognition, recommendation algorithms all exceed 20,000 RMB. For context, avg monthly salary in 37 major cities in China = 8,452 RMB. It’s not all rosy. According to one third-year computer vision major at a top university, “If you don't have a paper or a contest (result), it’s very difficult to find an algorithm (engineer) post.”

  • Suggestions from professors: Students should prioritize their own interests. They could also choose majors related to AI and specialize later. "In fact, the direction of the AI major is a little bit more narrow. One good option is for undergraduates to choose relatively broad majors, such as applied mathematics, statistics, automation, electronics, etc., to lay a good foundation first, and then to engage in artificial intelligence research at the graduate level," Huang Kaizhu, a professor at Xi'an Jiaotong-Liverpool University, said.

FULL TRANSLATION: "AI Major" has become a hot spot to apply for

Feature Translation #2: Ant Group’s IPO Prospectus

Context: “The US capital markets are being shunned by the largest initial public offering in history. This is an indirect result of the recent China-baiting by US politicians, led by Donald Trump,” writes Daniel Broby, a UK fintech expert.

Less than one month after officially announcing its intention to list on both the Hong Kong and Shanghai exchanges, Ant Group’s prospectus disclosed its financial status for the first time. In just the first half of 2020, net profits reached 21.9 billion RMB, which (impressively) far exceeded last year's annual performance. Let’s learn about Ant Group, formerly known as Ant Financial — via jiqizhineng (Synced).

Key Takeaways:

  • Over 60% of revenue comes from digital financial technology platforms. Among them, the source of profits is mainly from two important loan products, jiebei (“borrow it”)and huabei (“spend it”), especially huabei, which can provide Ant with at least several billion RMB in profits every year.

  • Before reading this, I would’ve sworn that Ant’s more well-known digital payments platform (Alipay) would be the largest source of revenues. It comes in second, but “There is no doubt that payments are the lifeblood of Ant Financial, and the valuation of Ant Group depends on the dominant position of Alipay. Industry insiders believe that the data accumulated by Alipay can be described as its ‘unique martial art,’ and the credit system based on this data can be described as Ant's most competitive product.”

  • A key point that the prospectus emphasizes, and the numbers bear out: Ant is a tech company. The company’s R&D spending averages around 8% of operating costs for past four year. Since 2017, the total number of employees of Ant Group has increased significantly, from 9,273 in 2017 to 16,660 as of June 30, 2020. Technical personnel accounted for 64% of the increase.

  • Why is the location of its listing important? Here’s where it gets even more interesting. Ant is listing in both HK and SH, but the company is only aiming to raise 48b RMB in SH, compared to 92b RMB in HK. This is a huge gap in the scale of fundraising between Shanghai and Hong Kong, and the latter will become the main battlefield for global investors.

  • The SH listing, on a new Sci-Tech innovation board (STAR Market), is still significant though. Recall, in ChinAI #94, there was speculation that Cloudwalk (a facial recognition startup) would list on STAR, which was established “after complaints that Chinese mega stars like Alibaba…chose to list in the US rather than at home,” to encourage investment in domestic tech companies and make it easier for mainland investors to trade in these companies. Ant would be the giant of the STAR Market. According to statistics from China Economic Weekly, on the first anniversary (took place earlier in 2020) of the opening of STAR, the total R&D expenditures of STAR’s 133 new stocks totaled 21.2 billion RMB. Ant Group's R&D expenditure in 2019 alone has reached half of all listed companies on STAR.

FULL TRANSLATION: Breaking Down Ant’s 500-Page Prospectus

ChinAI Links (Four to Forward)

Should-attend: Navigating US-China Tech Futures, Stanford Cyber Policy Center Online Seminar

I’ll be participating in this online discussion next Wed (10-11AM Pacific) with Andrew Grotto, Director of the Program on Geopolitics, Technology, Graham Webster, editor in chief of the Stanford–New America DigiChina Project, and Jennifer Pan, Assistant Professor of Communication in the Stanford Department of Communication. I plan to mostly be a fly-on-the-wall trying to soak up all the wisdom of the other three — though I will maybe occasionally lob a few hot takes @ these two quote tweets of the original event announcements:

How DARE we even consider the possibility of cooperation………

Should-read: Mapping U.S.-China Technology Decoupling

From an all-star team — Yan Luo, Samm Sacks, Naomi Wilson, and Abigail Coplin — DigiChina attempts to map the state of decoupling in the U.S.-China technology relationship. First systematic effort I’ve seen which collects Chinese and US government actions on export/import controls, cross-border data flows, supply chain security reviews, financial untangling, visa restrictions, encryption, website and app bans, and other efforts to reduce dependence on the other country. My quick reaction: this is a really important effort. If we did the same thing and mapped all the ways the US-China have become more technologically dependent on each other, what would be the net-outcome (decoupling or coupled for life?)

Should-read: Immigration Pathways and Plans of AI Talent

By Cahterine Aiken, James Dunham, and Remco Zwetsloot for CSET, a data brief: “To better understand immigration paths of the AI workforce, CSET surveyed recent PhD graduates from top-ranking AI programs at U.S. universities. This data brief offers takeaways — namely, that AI PhDs find the United States an appealing destination for study and work, and those working in the country plan to stay.”

Should-read: Cloud Security — A Primer for Policymakers

By Tim Maurer and Garrett Hinck for Carnegie, a very useful primer for policymakers on cloud computing and cloud security. One interesting takeaway is that the cloud can actually be a huge win for security but better understanding by policymakers is necessary to realize that. See, in particular, Figure 5 for a helpful mapping of the impact of cloud security risk vectors. Thanks to Ozzie Gooen of FHI for helping review this primer.

Thank you for reading and engaging.

These are Jeff Ding's (sometimes) weekly translations of Chinese-language musings on AI and related topics. Jeff is a PhD candidate in International Relations at the University of Oxford and a researcher at the Center for the Governance of AI at Oxford’s Future of Humanity Institute.

Check out the archive of all past issues here & please subscribe here to support ChinAI under a Guardian/Wikipedia-style tipping model (everyone gets the same content but those who can pay for a subscription will support access for all).

Any suggestions or feedback? Let me know at or on Twitter at @jjding99