ChinAI #107: Deconstructing Japan's Construction Industry

Plus, the first ever Eight to Exchange — catching up on last week's missed recs

Greetings from a world where an inland hurricane tore through Iowa and you probably didn’t hear about it

Fortunately, we in Iowa City were spared the worst of the storm, but there are some areas nearby that are really struggling. Some links to help out with donations if you can.

…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: Japanese Construction Robots Build Dams and Recruit Boston Dynamic Dogs

CONTEXT: One of my favorite platforms, the WeChat account jiqizhineng (Synced), reports on the progress of automation in Japan’s construction industry. Construction is one of the latest adopters of automation: robots can currently only handle 1% of the total construction engineering workload due to the diversity, sophistication and complexity of construction projects. This in-depth coverage of Japanese construction companies’ advances in automation is framed in a “look at how great the neighbor’s kid is doing in school” way. The conclusion states that automation in China’s construction industry lags far behind.

So…why are we talking about Japan’s construction industry in ChinAI? 1. It’s useful to see countries outside of the US as comparators for China’s AI advances. 2. While we’ve covered non-sexy AI (less visible, non-consumer-facing) mostly in the context of manufacturing, construction is facing its own Industry 4.0 overhaul. 3. There’s historical significance to this industry, as it was a key flashpoint in trade frictions between the US and Japan in the 1980s. Also, I found it interesting and came up with these reasons to justify the choice after the fact.


  • The drama of diffusion: The story doesn’t end when Boston Dynamic releases a flashy new innovation like its robot dog (“Spot”). How other companies adopt the innovation (the diffusion process) is just as important. For instance, Kajima Construction uses Spot to do tunnel inspection work. Spot is equipped with a 360-degree camera, which can take photos of the excavation site and inspect the instrument dials.

  • Graying of Japan’s construction workforce raises automation needs: 35% of workers are currently 55 years of age or older. Many of the giants have set up R&D labs focused on AI and robotics. The construction industry has also tried to encourage women to work on construction sites (image below is a hiring poster for female welders).

  • Automation vs. AI? Most of these construction robots, like Kajima’s concrete compactor follow pre-embedded programs. This is the case for most industrial robots as well. To get beyond 1% of robotic replacement for construction workload, they will need to be able to “perceive” the surrounding environment as well as learn “craftsmanship skills.”

  • Are giants necessary for introducing new technologies? Japan’s construction industry is dominated by five giant firms (Obayashi, Kajima, Shimizu, Taisei, and Takenaka). The article argues, “As far as the current situation in Japan is concerned, only when large construction companies take the lead in introducing advanced technologies such as robots and bring down related costs, then can we expect advanced technologies to be popularized throughout the construction industry.”

  • How does China compare? China has very few players (Country Garden, a Chinese property developer is mentioned). It’s reliant on imports for key robotics components and Chinese robot manufacturers neither have a strong background in the construction industry nor have accumulated robotics application technologies.

FULL TRANSLATION: Japanese Construction Robots Build Dams and Recruit Boston Dynamic Dogs

ChinAI (Eight to Exchange)

Must-read: The human cost of a WeChat ban, severing a hundred million ties

Here’s the excerpt about how a WeChat ban would affect me personally and professionally:

Jeffrey Ding, an American researcher at Oxford University who studies China’s AI strategy, also uses WeChat to keep in touch with family and friends as well as scholars. Over 80% of the blogs and documents he translates as part of his work are sourced from WeChat public accounts and the WeChat messaging groups that include Chinese researchers.

“Alternatives do exist, and if I have to adjust I will,” he says. For example, he’ll switch back to phone calls to replace his weekly video chats with his grandma. But while he’s more confident he can find substitutions for his closest relationships, he worries about the weaker connections he’ll lose in his network. “We shouldn't discount the significance of these ‘thinner’ ties,” he says. “Sometimes the thinnest of ties can lead to much deeper understanding and open up doors when the opportunity arises.”

Must-read: Shock of the Old by David Edgerton

This week’s feature article discusses how Kajima’s work on the Oitagawa Dam does not end when the dam is built— the maintenance work begins immediately. They use drones to perform laser measurements of the site and also to monitor/manage the dam. That Shock of the Old devotes an entire chapter to “Maintenance” gives you a sense of how it tells the history of technology in a unique way. It’s probably the book that has most influenced my outlook on thinking about technology.

Should-read: Rhetorical Frames of an AI Arms Race

Very cool CSET issue brief by Andrew Imbrie, James Dunham, Rebecca Gelles, and Catherine Aiken. They searched more than 4,000 English-language articles over the 2012 to 2019 period from the global news outlet Reuters, the U.S.-based Defense One, and Foreign Affairs and identified references to “AI competition.” Key takeaways:

  • The rise of AI coverage is just astronomical. The first mention of AI in Defense One was in 2014! If I’m interpreting the report correctly, that means from 2012 to 2013 there were no articles about AI in Defense One. In 2019, Defense One published 215 such articles.

  • Since 2012, a growing number of articles in the three news sources have included the competition frame, but the prevalence of the frame as a proportion of all AI articles peaked in 2015. Reporting on AI has become more sophisticated and diverse.

  • *I think this brief gives a positive indicator that coverage about AI is maturing — the diversity of frames is especially important. Still, I’d be curious if we see a decline in the AI competition frame if the sample were limited to articles that mention China.

Should-apply: FHI Research Scholars Programme

The Future of Humanity Institute’s Research Scholars Programme is a selective, two-year research programme, with lots of latitude for exploration as well as significant training and support elements. This year we will offer roughly eight salaried positions to early-career researchers who aim to address the big-picture questions critical to humanity’s long-term flourishing.

Applications for the spring 2021 cohort are now open. Deadline: 12:00GMT on 14th September. Apply here for the Research Scholar position, or here for the Senior Research Scholar position.

Should-listen: China already has your data. Trump’s TikTok and WeChat bans can’t stop that

Some needed perspective from Aynne Kokas in the Washington Post — argues that the executive orders on WeChat and TikTok are merely window dressing without comprehensive data security regulation. The piece certainly sticks the landing: “Data exfiltration from the United States to China is systemic, and presents a pressing national security issue. Keeping consumer data safe would require comprehensive oversight of, and transparency from, all firms operating in the United States. Spotty executive orders, half-written in response to stunts by K-pop fans, won’t cut it.”

Should-listen: WeChats from the Future (NPR’s Rough Translation)

A story about the time lag between the arrival of the coronavirus in two different nations, and how that played out in a marriage. Rough Translation, one of my favorite podcasts, explores the story of a Chinese American wife (Liying) and her husband who live in Connecticut. Liying, originally from Wuhan, is more connected to online chat groups of local Chinese-Americans and gets much more worried than her husband about a mysterious virus that hits her hometown.

Should-read: China Won’t Win the Race for AI Dominance

Two Oxford researchers Carl Benedikt Frey and Michael Osborne argue this case in Foreign Affairs, analogizing U.S.-China competition in AI with U.S.-Japan competition in computers back in the 1980s. GovAI is planning a webinar on this topic with Carl and Michael, so stay tuned!

Should-listen: In Machines We Trust

Very excited about this new podcast from MIT Tech Review and host Jennifer Strong, who previously hosted WSJ’s The Future of Everything podcast. The first four episodes cover facial recognition and policing.

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