By Yina Moe-Lange
The previous blog post discussed the pace of AI and came to the conclusion that there is no precise consensus among experts of the speed at which AI will evolve. While there was great variation among the experts, there was an interesting difference in the between answers from Asian researchers and North American researchers. The Oxford and Yale study found that Asians predict High-level Machine Intelligence (HLMI) in 30 years compared to 74 years for North Americans. This difference of 44 years is pretty significant as it also the average length of a career, meaning that Asians predict HLMI by then end of this generation of careers, while the Americans expect that plus one more generation. The two countries with the most respondents China and the US, show an even bigger difference with predictions of 28 years and 76 years respectively.
In the figure below, you can see the trajectory of HLMI probability by region. With the red line representing Asia quite a bit higher up than the other regions, and with North America the lowest.
China’s rapidly evolving commitment to AI can be seen through other channels as well. A number of journal articles published out of China has caught up to and passed the US. The Obama Administration put out a strategy plan on AI research and development in October 2016. They looked into a number of journal articles published in different countries around the world. Up until 2009, all countries were publishing about the same, low number of articles. Then in 2010, the US starts publishing more than its peer countries with China not too far behind. Both countries, since then have had exponential growth in output but in 2014, China recently surpassed the US and now lies solidly ahead of them. Below, the figure shows the number of journal articles mentioning deep learning or deep neural network, that have been cited at least once by country of origin.
China has been heavily investing in AI and Machine Learning companies in the US and around the world. Although China is a big investor in AI projects in the US, the US government has begun to “rethink” these investments, citing worries about China’s access to AI technology in the US. Unfortunately, regulation of these investments could slow down the US’ growth in the AI sector due to a significant drop in overall funding. It will be interesting to see if there is a follow-through on the regulation and if there is, if there is a significant drop in research and growth of AI in the North American region.
Once last interesting trend. When using Google Trends for Machine Learning, you can see that the regions with the highest interest in Machine Learning are all in Asia. The US is #11 on this list.
Drawing from various research and studies, it is possible to put together a rough idea of where the two regions are in terms of research and assess their current trajectories. This picture can change quite a bit as government funding catches up and more and more large companies join the AI market. The US’ announcement regarding limiting Chinese investment in AI technology is concerning as it could have an unintended negative impact on the growth of AI and AI development in the US.