16 January 2018 - 09:30
AI, a two-horse race
The rivalry between China and the USA in the field of Artificial Intelligence (AI) is becoming crucially important
When it comes to Artificial Intelligence, the competition between the United States and China has already begun to affect investments in the sector: the figures and results are impressive, but as yet there is no winner
In July 2017 the Beijing government published a planning document in which it set the goal of becoming the global superpower in AI by 2025, overtaking the Americans, who are currently leading the way. Reuters reported that the Chinese aim to grow the country’s core AI industries to a value of 22 billion dollars by 2020 and 60 billion by 2025. “Artificial intelligence has become the new focus of international competition. (…) The major developed countries in the world regard the development of artificial intelligence as a major strategy to enhance their national competitiveness and safeguard their national security,” reads a Chinese government publication quoted by CNN.
The US authorities’ lack of ambition has been noted by a number of experts, including those of the Harvard Kennedy School’s Belfer Center for Science and International Affairs, which recently published a study of Artificial Intelligence and national security. The research suggested that the USA is at risk of ceding its current advantage in military and non-military AI.
Today, of the 2542 companies operating in this sector on a global level, 42% are based in the United States and 23% in China. The figures for investments highlight the duopoly that has emerged. According to the website Technode, of the 25 billion euros invested in the sector since 1999, 51% of this sum has ended up in the United States and 33.18% in China, the rest of the world was left with a mere 15.73%.
The big Chinese companies are preparing to take on their American rivals. Baidu, the country’s leading search engine has invested significantly in AI and already boasts two enviable developments. The first is DuerOS, a highly advanced vocal recognition system that enables interaction between man and machine: the former issues commands and the latter obeys. With DuerOS, Baidu is directly challenging the heavyweight American products such as Amazon’s Alexa, Google’s Assistant, Apple’s Siri and Microsoft’s Cortana. Furthermore, with the Apollo software the company has made an assertive entrance into the race to develop self-driving cars. Apollo was launched in 2013 with an investment of 1.5 billion dollars. The results so far indicate that the software is being used by six thousand programmers. The initial experiments are now due to take place in some simple driving situations. By July 2018, Baidu is scheduled to have begun production and trials of its own fully autonomous minibus, while by 2020 the company’s self-driving car is due to become a reality.
It is precisely in this field that Baidu will collaborate with the Ministry of Science, which has selected the company as part of the public-private “dream team” assembled to scale the heights of AI. The other companies involved are Alibaba, Tencent and iFlyTek. Alibaba is set to work on solutions to improve life in cities, Tencent will deal with computerization and diagnosis while iFlyTek will be working in the field of vocal recognition.
Though the plans might seem ambitious, there is no lack of resources and the public-private collaboration is intense. The United States continue to be in the driving seat, also due to the fact that American AI, in difference to its Chinese competition, has a substantial history.
The only sector of AI in which the Chinese seem to be ahead is that of intelligent robotics.
The most notable disparity, however, can be seen in the respective talent pools: in the United States there are 78 thousand AI specialists, in China 39200. Much depends on the universities; Chinese institutions are not yet at the level of their American counterparts. This would indicate that in spite of the apprehension that exists in the United States, Beijing’s progress is set to take longer than anticipated. Two particular examples are worth noting: Google’s Alphazero supercomputer, which in just four hours was able to master the rules of chess and immediately outclass Stockfish, the champion AI chess program. The other example is NASA’s discovery of another world, a planet similar to our own but extremely far away, made possible thanks to the technology provided by Google.