I have received a number of emails in response to my recent post on China’s artificial intelligence plan. Many who wrote me seek to reduce China’s plan to the following simple, three step process:
- Catch up in AI by 2020.
- Learn to make some basic products by 2025.
- Lead the world by 2030.
This does not accurately summarize the plan, though it is how much of the English language media describe it. The full PRC AI plan is set out in 35 pages of dense, jargon heavy, Chinese bureaucratic prose. I will be doing a series of blog posts seeking to explain the full plan. My first post, China’s Artificial Intelligence Plan — Stage 1, dealt only with stage one, and as you can see from that post, stage one is not written as “catch up.” Stage one is a full-on plan to continue developing technologies with which Chinese companies are already working. I note also that manufacturing automation robotics is not featured. On the robotics side, the emphasis is on service robots.
Note also that the Chinese companies are already way ahead of this plan. They are not waiting around for guidance from the government on their AI projects; they are moving ahead full speed. In general, Chinese companies are succeeding most with the software/network based applications of AI. This is the focus of the Baidu research center in Silicon Valley. They are not doing as well with mechanical devices such as robotics and smart vehicles and sensor based IoT devices. However, they know that and they are making strong efforts to advances in these areas. We touch on this a bit in China IP Challenges for Automotive Suppliers. One of the areas on which many Chinese companies are focusing is on human/AI interaction and they are having good success with that in the field of medical imaging and diagnosis.
There is little doubt that part of China’s AI strategy involves acquiring technology and then selling in back into the developed market from which it came. This has been and still is the strategy of businesses in pretty much every developing country. The U.S. followed this approach during the entire 19th and early 20th centuries. Japan and Korea and Taiwan did it with great success in the post WWII era. China and India are now moving into that phase. That is how technical progress is made and we write about to guard against this sort of IP appropriation nearly every week. See How To Give Away Your IP In China, How to Give Away Your IP in China Without Realizing It and China and the Internet of Things and How to Destroy Your Own Company.
The real question is whether this strategy will work in the AI era In general, Chinese companies are not good at working on their own to appropriate foreign technology. They prefer to enter into a manufacturing or joint venture arrangement where they convince the foreign entity to teach them how to use the technology. See China Joint Ventures: Keeping Your Friends Close and Your IP Closer. Then, later, they appropriate the technology and sell the cheaper product back into the same market. This is what Chinese companies did with high speed rail and with the Russian designed fighter jet and this is what they are trying to do it now with commercial aircraft. They will undoubtedly seek to do the same thing with AI and robotics. See China-US Trade Wars and the IP Elephant in the Room.
Will China succeed it purloining AI IP? It depends on a couple of factors. First, if the technology is protected by patent and copyright and trade secrecy, then they cannot sell into the markets where those IP protections exist. This would mean that North America and the EU would be closed to them, at least during the period where the IP protections are in place. Second, can Chinese companies master the technology to point where they can really compete? Normally, the Chinese companies simply clone the product and then seek to compete solely on the basis of price. For some products, this works. For more sophisticated IoT, AI, “smart” products, the success rate of the Chinese companies has been low. How many U.S. consumers get excited about the purchase of a Lenovo computer or a Xiaomi cell phone? How many U.S. customers are interested in buying PRC knock offs of virtual reality headsets? Not many. Price is not the significant issue for these more technically sophisticated products. When Chinese companies cannot compete on price, they traditionally don’t know what to do. There are many programs in place in China focused on changing this “price is the only issue” mindset. So far, progress has been sporadic at best.
However, without regard to whether China can succeed with its AI program, it is clear that appropriating foreign AI technology is the goal of most Chinese companies operating in this sector. For that reason, foreign entities that work with Chinese companies need to be aware of the significant risk and take the necessary steps to protect themselves. There are many ways to do this, using a mix of IP registrations and carefully drafted agreements. See China Contracts: Make Them Enforceable or Don’t Bother. This is what the China lawyers at my firm focus on and this is the issue we discuss most often on this blog. Stay tuned.