How is China able to catch up with US in AI research and output ?

The whole world is fully aware of the potential of Artificial Intelligence and every scientifically advanced country is working hard to establish itself as a front runner in developing AI technologies. US traditionally has been the country that has been leading the world with almost 2000+ AI companies researching, experimenting and developing AI products for the world. Click here for a good introduction to the field of Artificial Intelligence from IBM.

Of late, China has overtaken US in AI research producing about 54% more papers at around 37,343 compared to the US with approximately 24,000 papers in 2017. (click here) The way China is developing in the global field as regards Artificial Intelligence is really astonishing for the world. 

China which was a non scientific entity especially in the Computer Science Engineering area about twenty five years ago, in just twenty years has developed fast and is presently the world leader in the number of publications and patents in the area by a good margin of 20% over the former leading country US.

How did this happen ? Click here (same as highlighted above) for an interesting Harvard Business Review article that describes how China has brought about this change. 

Catch up cycles happen when latecomers because of modern technology, favourable government policies and market environments are able to catch up with incumbents. 

China has caught up with the US especially in AI technology and applications. Three reasons are cited for this change. These reasons are a combination of demographic, scientific reasons and nature of the AI field besides favourable policy decisions of the Chinese government in promoting cutting edge AI research in Chinese companies. The far sighted vision of the Chinese political leadership cannot be discounted.

Firstly AI technology is not protected and is available largely in public domain as many researchers are working and publishing their results. This has enabled fast percolation of technology across applications and fast adoption.The Open Science nature of AI has helped in the fast spread of knowledge. 

Secondly data and Computer Science knowledge, talent and innovation in the AI field has got more impact on growth than patents that capture knowledge to pockets of usage. The conventional pharma industry by holding patents can give companies the edge over newcomers and competitors and kills innovation.

In the weak AI area, one makes an AI product puts it in public domain or use to generate data. This data is captured and using machine learning algorithms is analysed leading to refined product usage through data. The cycle continues. Sharing of the AI product to the public as early as possible is thus necessary for the early success of the product.

Thirdly China's large population and markets gives the AI companies the unique opportunity to handle large amounts of data and assemble huge databases. For example, Didi, the equivalent of the car hailing app Uber in China, on a daily basis handles 70TB (70,000 gigabytes) of data, with 9 billion routes being planned daily and almost 1000 car requests happening every second. The volume of data generated is so vast, the machine learning algorithm can learn by itself to take improved decisions after every cycle.

In the near future we can expect Chinese AI products and offerings overtake US, catapulting China to the numero uno status in the field. The Communist dictatorial leadership in China and it's wholly whimsical ways in dealing with dissent, competition and openness is already a matter of concern for the global scientific community. Hope China will come out open and public in it's AI policy for the whole world to be sure that AI is not being used in undesired ways for world domination in future.

George.