After a detailed study of the current development of Artificial Intelligence (AI) and its market and a careful reflection of cross correlation between AI and IP in the new period, we will probably be able to see their future development directions and tracks at the crossroad where they meet, so as to hold the fate and well-being of human in our own hands.
AI, as the name suggests, can be interpreted as a technology that enables the computer to realistically imitate the human way of thinking and thinking process and have the unique human abilities (intelligence) including learning, feedback, communication and decision-making by means of design by human and the interaction between human and the computer. AI is usually regarded as a branch of computer science, but it also involves a lot of logic, mathematics, linguistics, psychology and other disciplines. As early in 1950s, A.M.Turing, Father of Computer Science, proposed the famous Turing Test: if a computer is able to answer questions proposed by human in a way that cannot be told from human, then such computer may be considered intelligent. Today, to give the computer the ability to accomplish what requires human intelligence in different fields has become the most important goal in the research of AI.
Since the concept of AI was first born at the Dartmouth Conference in 1956, the research, investment, financing, enterprise construction and social concerns in the field of AI have always been in a state of rapid growth. In recent years, AI has been increasingly becoming one of the important engines to push technological progress and promote a new-round technological revolution and industrial revolution. As intelligence technologies including speech recognition, face recognition and smart home gradually go to people's daily lives, followed by hot events such as AlphaGo's winning streak over Lee Se-dol and Ke Jie in human-computer competitions, AI, a complicated, difficult technical field, seems to have become reachable in the eyes of the world. On March 5, 2017, AI was first mentioned in the government report. On July 8, the State Council released the Development Planning for the New- Generation Artificial Intelligence in a move to officially include the AIoriented technological and economic development in the national strategic level system layout. All these events have shown that the era of AI is coming unstoppable.
Today, to IP professionals, they are also closely related to the development and progress of AI technologies. On the one hand, the powerful capacity of AI in data collection and analysis can greatly enhance the work efficiency in patent retrieval, patent review and infringement research etc.; on the other hand, with the rapid development of AI, the research on, and applications for, AI technology patents are in rapid growth throughout the world. Due to its own particularity, AI also involves a lot of issues with respect to the protection of IP and has resulted in a reflection in the industry concerning the concept of IP and its operation update in the new period, such as how to protect the intellectual property of data and how to determine the copyright owner of works created with AI. Therefore, it is not that easy to define the relationship between AI and IP.
Current development of AI globally: technology, market and research
After more than sixty years’ development, the AI technology has developed a number of major subfields, including deep learning/ machine learning, computer vision, natural language processing, context awareness and computing, intelligent robots etc. Today, AI is mainly used in multiple industries including finance, education, security, e-commerce retail, medical health, and personal assistance, as well as in other industries. AI products such as Apple's Siri, IBM's Watson and other various face recognition technologies have been widely used worldwide. In the global context, AI plays an important role in improving the production and service efficiency, accelerating industrial upgrading and promoting an industrial revolution.
According to the Wuzhen Index: Annual Report on Global Artificial Intelligence (2016), the global AI market was 168.39 billion Yuan as of 2015 and is expected to reach 270 billion Yuan in 2018; China's AI market had exceeded 10 billion Yuan as of 2016 and is expected to reach 34.4 billion Yuan in 2019.
The AI market has broad development prospects, attracting more and more enterprises to enter this business. At present, the global AI enterprises are very unevenly distributed, and most are found in the US (2905), China (709) and the UK (366), accounting for 65.73% of the world's total number of AI enterprises. Since 2000, the number of global new AI enterprises annually has shown a linear growth trend; since 2013, mergers and acquisitions in the global AI field have also begun to increase year by year. Google, IBM, and SoundHound are leading and reputable in the AI industry worldwide. In China, Beijing (30.74%), Guangdong (21.60%), Shanghai (15.17%), Zhejiang (7.04%) and Jiangsu (6.03%) together have gained the vast majority of China’s domestic AI market share, together up to 80.58% in terms of the total number of AI enterprises. Since 2000, the annual growth of China's AI enterprises has increased year by year, and especially the increased AI enterprises from 2014 to 2016 had accounted for 55.38% of the total cumulative number of China's AI enterprises. Baidu, Tencent, Alibaba, and iFLYTEK are AI leaders in China.
In addition to the growth in the number of enterprises, the rapid expansion of investment and financing in AI also shows the vigorous vitality of the AI market. Similar to the distribution of global AI enterprises, the US (USD 17.912 billion), China (USD 2.572 billion) and the UK (USD 816 million) are the top three for investment and financing in AI. Over the past ten years, global AI financing has centered on the seed round. In terms of the number of institutional investors, the US (900) ranks No. 1, leaving far behind the UK (96) and China (43); in terms of the number of investments, the US (3454) also leads way ahead when compared with the UK (274) and China (146). In China, Beijing, Guangdong, Jiangsu, Zhejiang and Shanghai, where AI enterprises are gathered, attract most of AI financing, together up to 91.45%, including 50.22% of financing in Beijing's AI enterprises.
As key locations of scientific researches, universities have an irreplaceable role in the development and promotion of AI technologies. According to Global Universities of AI Influence in 2017 from the Wuzhen Index: Annual Report on Global Artificial Intelligence (2017), MIT, Stanford University, and Carnegie Mellon University rank top three; the top 30 universities are mostly from the US, followed by Europe, and include only 2 Asian universities from Israel. In the List of China's Universities of AI Strength published by Wuzhen Institute, Tsinghua University, Peking University, and Zhejiang University rank top three.
In a word, since the 21st Century, global AI has made great progress in technology, market and scientific research, and has maintained the trend of accelerated development on the whole. It should be noted that now AI development in the world has shown a unipolar pattern: the US has leading edges, followed by Europe and China, while AI development in other countries or regions needs to be improved. AI still has great room for development in the world. As a rising star China has made remarkable achievements in this field in recent years, with its development rate higher than that of the top countries within the industry including the US. Of course, when it comes to various indexes, there is still a big gap between China's AI and the world's top level and the AI research strength remains to be strengthened. In China, AI also shows very uneven development, concentrated in a few provinces and cities. China's AI has a lot of room for improvement in terms of technological R & D, market development and regional coordinated development.
Thriving: overview of global AI patents
Among the various types of IP rights, there is no doubt that patents are the most closely related to AI technologies and market. Since the 21st Century, the rapid development of AI technologies and the entry into AI field by enterprises have promoted the growth in the number of AI-related patents.
According to the Wuzhen Index: Annual Report on Global Artificial Intelligence (2016), as of 2016, there had been more than 77500 AI patents included in the global patent database, and the countries with the most PCT applications/ grants of AI patents were the US (26891), China (15745) and Japan (14604), together up to 73.85%. In terms of sub-fields, global AI patents mainly fall into intelligent robot (32.5%), speech recognition (20.8%) and neural network (20.7%). The geographical distribution pattern of China 's AI patent applications is basically the same as the distribution pattern of AI enterprises. According to the data on the number of applied/granted patents as of 2016, Beijing (7841), Jiangsu (6675), Guangdong (5261) and Shanghai (4222) were currently China's four largest gathering centers of AI patents. In terms of sub-fields, China's AI patents mainly fall into intelligent robot (38.3%), neural network (17.9%) and image recognition (10.4%).
At present, the number of global AI patents is growing at a remarkable rate, especially in China. According to statistics, in the past five years, the average annual growth rate of China's AI patents has reached 43%, about twice that of the US (21.7%). So, China's AI enterprises are catching up with the world's top level in terms of AI research capacity and the protection of IP.
Promotion and impact of AI on the IP industry
The rapid development of AI technologies has improved the data collection and analysis capacity of the computer in line with human needs to a new level. For patent retrieval, patent review, infringement research and other IP related fields. that require a large amount of data, the progress in AI technologies is undoubtedly a great news. It will benefit both related government agencies and enterprises which have the intention of the research and patent layout.
Patent strategy establishment, patent layout advancement and cross-department coordination in enterprises will rely on the deep understanding, analysis and search of patent information. Today, as there is abundant dispersed patent information, it is becoming more and more difficult for enterprises to acquire, process and distribute patent information, quite different from the past when patent information was monopolized by the IP departments of few enterprises. Under the new situation, enterprises should first have a deeper understanding of the nature of information in order to effectively and efficiently search patent information.
In fact, AI should act as a collaborator rather than a substitute for human intelligence in the path of patent data enterprise decision making. With its strong computing power, deep learning ability and multilayer neural network, AI can help with the simultaneous quick search in multiple databases, in order to instantly find the similarity, frequency and similar phrases of a concept and sort them in high quality according to correlation. So it has prominent advantages in the quick collection of data, integration of nonsignificant data, and extraction of key information. In addition, AI often supports short learning curves, able to adapt to the needs of mature users and help beginners get started quickly, in a way to reduce the artificial learning costs; when it comes to application development and decision making based on patent information, human brain still has irreplaceable advantages. Compared with the traditional way in which a large amount of repeated patent data and information is processed manually, it is possible to process patent information with the help of AI to bring an amazing promotion in efficiency and accelerate the monetization of patents. More importantly, AI can also help enterprises fully understand and use the innovation factors around the world to achieve forward-looking investment and reasonable allocation of resources, so as to provide a powerful boost for the long-term development of enterprises. In view of this, it has become popular among more and more enterprises to update the traditional patent strategy with the help of AI for the research of patent infringement in the new way to process and utilize patent information.
By improving the efficiency in processing and utilizing patent information, AI technologies can help enterprises make decisions and also play its important role in patent review. In addition, the collection and review of patent information often relies on accurate patent translation; as an important branch of the AI technology, neural network and natural language processing technology distinguish in improving the efficiency and quality of machine translation. The large amount of repeated, mechanical translation and review work which requires translators and reviewers to conduct now can be gradually processed by AI, where IP-related government departments will surely see the shortcut to improve efficiency and reduce costs.
Similar to the conditions in other industries, while bringing great convenience to the IP industry, AI has inevitably caused concerns from some practitioners. As mentioned earlier, since AI features low cost and high efficiency in patent review and translation, it is likely to result in a reduction of reviewers and translators in the future. Meanwhile, the AI technology has been tried for patent application writing, which may threaten the positions of patent agents. On June 5, 2016, the Japan Patent Office accepted a patent application allegedly to be the world's first one written by AI, which was reported to have functions including automatic generation of application documents, automatic access to IPDL (Industrial Property Digital Library), checking of repeated applications, and automatic online applications at a daily frequency of more than 60,000. Obviously, once such AI technologies become mature, many patent agents seem to be in danger of losing their jobs.
To sum up, AI technologies have entered the field of IP especially the patent field, and this trend is persistent and irreversible. No matter to what extent the concerns of intellectual property practitioners are to become true in the future, that AI technologies will replace human positions in some categorized or mechanized work is a general trend. Of course, the unique advantages of human intelligence in abstract thinking and reasoning are still out of reach for AI in foreseeable expectations. Looking forward to the future, how to achieve the improved collaboration between AI and human intelligence may be a more meaningful question than "if AI can completely replace human wisdom".
New issues of IP in the Era of AI
As AI technologies are gradually entering every corner of human production and life, we seem to have heard the footsteps of the fourth industrial revolution in full swing. The industrial revolution relies on innovations, and innovations rely on the protection of IP. Therefore, in the face of the rapid AI development, we also need to ask: today, can the human knowledge and understanding of IP and the IP protection systems in different countries effectively protect AI?
To answer the above question, first it is not possible to avoid the IP protection of data. AI, Internet of Things, big data and other new technologies are all built on the basis of data collection and analysis. Data has become an important source of enterprise competitiveness. However, while people are gradually aware that data should be subject to clear, effective protection of IP, IP-related systems and laws and regulations of all countries are still lacking in this regard, resulting in concerns from a lot of data practitioners. And given the characteristics of data in its form, content and nature, the scope and method of its IP protection as well as the advantages and disadvantages in a separate legislation for protection and/or modification of existing laws for protection are also problems before the IP industry. To this end, Imamura Wataru, director of Policy Planning and Research Division at Japan Patent Office pointed out that the IP protection of data should start from a legal point of view. A separate legislation may not be necessary, but the modification of the existing patent laws in a timely manner should be an inevitable prerequisite to meet the needs of new technologies such as AI.
Another new IP issue brought by AI technologies is AI inventions and the involved IP protection. In May, 2017, the collection of poems composed by Microsoft XiaoIce (an AI robot) was published and claimed to be the first collection of poems created entirely by AI in the human history. The publication of this collection has brought again to the public the long controversial question "whether works created by AI should be copyrighted".
Different from the protection of IP of AI technologies, the protection of AI inventions and works often involve technological ethics. Should AI inventions and works be subject to IP protection? If "yes", then should such kind of IPbe owned by the technology inventor or AI itself? At present, answers to this question are more from the legal point of view rather than from the technological point of view. This year, during discussion of copyright protection of AI works in Japan's academic circles and industry, attribution of IP to the technology inventor rather than AI itself based on originality pursuit gained more recognition. However, with the further development of AI in the field of creative labor, maybe someday in the future, human will eventually break the traditional thinking that only a natural person may be the subject of IP.
Conclusion
At present, the rapid development of AI technologies and the full expansion of markets will gradually make the past fantasies of human production and life come true. For the global IP field, AI technologies will bring major changes, creating both opportunities and challenges. More importantly, in the unknown future, AI may become a new member in the human world, which is likely to lead to a fundamental change in the concept, system and operation of IP for the human. We have come to the crossroad between AI and IP, where various factors converge. It is not clear as to where is the road to the future.
The uncertainty of the future does not necessarily mean the confusion of human. After a detailed study of the current development of AI and its market and a careful reflection of cross correlation between AI and IP in the new period, we will probably be able to see their future development directions and tracks at the crossroad where they meet, so as to hold the fate and well-being of human in our own hands.
(Translated by Liu Lei)
京公网安备 11010802020770号