2024
600 million users and 200 large models. How generative artificial intelligence is developing in China
As of October 2024, there are more than 600 million users of generative artificial intelligence (GENI) services in China. At the same time, over 200 large AI models were launched. The technology has become the "engine" of industrial transformation, as stated in the materials of the Ministry of Industry and Information Technologies of the PRC (MIIT), published on October 12, 2024.
The popularity of Genia-based applications and services in China is growing rapidly. Numerous factors contribute to this, including the acceleration of technological innovation, as well as additional advantages in the fields of production, finance and healthcare. GeneAI integrates with other emerging industries and products such as humanoid robots, smartphones, computers, self-driving cars, etc.
 | The scale of the artificial intelligence industry in China continues to expand, and the number of enterprises in the relevant area has exceeded 4,500, says Zhao Zhiguo, chief engineer at MIIT. |  |
At the same time, Li Baiyang, assistant professor of intelligence research at Nanjing University, notes that previously an estimate of the number of users, including mobile ones, was used as indicators of Internet development. But in the context of the rapid transformation of the global IT industry, the number of registered users of large AI models has become a new indicator. As of October 2024, the Chinese artificial intelligence industry has taken a leading position in both large open-source and closed-source language models, Bayyan said. The country's authorities intend to encourage the use of generative AI technologies in industries such as manufacturing, agriculture, education and health.[1]
Chinese driving schools teach future drivers in cars with autopilot replacing instructor
In early August 2024, it became known that Chinese driving schools began to use artificial intelligence to train future drivers. For this, robomobiles equipped with a special autopilot system are used.
According to the South China Morning Post, an innovative AI platform has been introduced in a Beijing driving school. Beginners are trained on cars without the actual presence of an instructor in the car - the process is monitored by a neural network. With the help of a special course, the robomobile consistently trains the driver in the necessary actions on the road. After that, the student needs to perform test runs to consolidate the material.
Normal instructors are on site or in the control room. The AI platform allows one person to control the learning process of several novice drivers at once. This reduces the cost of courses. In addition, the throughput of driving schools is increased. It is claimed that up to 20 cars can be at the training ground at the same time, followed by only three instructors.
The AI mentor issues instructions to students and tells them where they should turn or stop. In the cabin of the vehicle there is a display on which a map of the site is displayed. The driver can see the markings and certain instructions. Safety is monitored by various sensors, which, among other things, prevent collisions and hits on any obstacles. Cars can follow each other with the necessary distance.
| | Beijing's high-tech driving school approach reflects a broader trend in China as it seeks to become the world's leading manufacturer of self-driving vehicles, the South China Morning Post said.[2] | |
US imposes ban on the supply of AI memory chips to China
On July 31, 2024, it became known that the US authorities were expanding the list of sanctions measures against China. In particular, a ban is being introduced on the supply of high-performance AI memory chips and equipment for their production to the PRC. Read more here
US companies banned from investing in AI and quantum computing in China
At the end of June 2024, it became known that the US Treasury Department had prepared new requirements prohibiting American companies and individuals from investing in high-tech Chinese enterprises. We are talking about areas such as artificial intelligence, quantum computing and advanced semiconductor chips.
The goal of the initiative is to prevent the use of American developments and funds to strengthen the IT potential of the PRC. This is being done to ensure US national security. Investors who want to invest in Chinese enterprises will have to provide the American authorities with detailed information about the planned transactions.
China's Commerce Ministry says the new requirements represent an "excess of permissible norms under the guise of national security." US policy is said to undermine the international economic and trade order, disrupting the security and stability of global supply chains. This can have a negative impact on the IT industry as a whole.
| | China expresses serious concern and strong disagreement with the new requirements and reserves the right to take appropriate measures, the PRC Ministry of Commerce said. | |
According to the former Vice Minister of Commerce of China Wei Jianguo (Wei Jianguo), the US government continues to tighten control over exports and investments in the PRC. The new restrictions, as noted, violate normal trade and investment activities between Chinese and American companies, complicate bilateral technology exchange and undermine the global IT ecosystem.
| | The move is the latest evidence that the U.S. government is using all means to curb China's technological growth, regardless of the complexities it will cause American companies, Jianguo says.[3] | |
US shuts down China's access to AI training in American cloud systems
At the end of January 2024, it became known that the US Department of Commerce is introducing new restrictive measures for Chinese companies regarding cloud technologies. The agency will close access to AI training in American cloud systems for customers from the PRC.
Washington is concerned that Chinese companies, despite the current sanctions restrictions, can access computing resources based on the latest GPUs through cloud providers such as Amazon Web Services (AWS), Microsoft Azure and Google Cloud. Thus, PRC customers can use American cloud platforms to train their own large language models (LLMs). At the same time, China's development of artificial intelligence and other technologies of a new generation is the main concern of the administration of President Joe Biden, which sees Beijing as its main strategic competitor.
| | We begin the process of requiring US cloud providers to tell us every time a foreign company uses their capacity to train their own LLMs, says US Secretary of Commerce Gina Raimondo. | |
In general, the US authorities reportedly intend to block all opportunities for China to gain access to American LLMs or to create their own AI models using the computing power of American cloud platforms. The difficulty lies in the fact that cloud service providers do not physically transfer any goods outside the United States, and therefore it is problematic to regulate their activities based on export control rules. At the same time, cloud companies themselves declare that restrictions on interaction with foreign users can put American firms at a disadvantage.[4]
2023
Market growth of 13.9% to $79.5 billion
In 2023, the volume of the main industry artificial intelligence China reached $79.5 billion (578.4 billion yuan). This is 13.9% more than in the previous year, when the costs in this area were estimated at $69 billion. The PRC authorities intend to introduce more than 50 new national and industry standards in the field of AI by 2026. This is stated in the document of the Ministry of Industry and, information technology China with which TAdviser I got acquainted in early September 2024.
It is noted that by the end of 2023 in the PRC, more than 4,000 companies and organizations conducted activities related to AI technologies. Vice President of the Chinese Academy of Sciences Wu Zhaohui says that various AI solutions will have a practical impact on all key industries, including manufacturing, transport, education and healthcare. The country is actively expanding smart production sites - both at the national and provincial levels.
The indicator of the introduction of generative AI (Genii) by Chinese enterprises in 2023 reached 15%, and the total financial effect of the use of such applications and services is estimated at 14.4 trillion yuan ($2.02 trillion). According to McKinsey, generative AI can additionally bring economic benefits to China in the amount of up to $2 trillion a year. The rapid development of the local Genii market and the growing adoption of technologies by enterprises will have a significant impact on the overall digital economy, will stimulate innovation, increase efficiency and will lead to increased competitiveness of Chinese companies.
| | Artificial intelligence technologies are rapidly changing human life - from humanoid robots to autonomous vehicles. Such products can bring transformative power to regional economic growth and change the landscape of global competition, the publication says news.cn.[5] | |
In China, the number of AI-related patents for the year increased by 40% to 378 thousand units
As of the end of 2023, the number of existing patents for inventions in the field of artificial intelligence in China reached approximately 378 thousand. This is 40% more compared to 2022, when the indicator was estimated at 270 thousand. Such data were published on July 29, 2024 by the State Administration for Intellectual Property of the PRC.
The Chinese AI industry is rapidly developing. It is noted that the growth rate of the number of patents in this area in the PRC is on average 1.4 times higher than global values. AI is one of the key directions of the country's digital economy. In 2023, the main sectors of the digital economy provided approximately 10%. GDP China By the end of 2023, a total of 155 thousand Chinese enterprises had patents for inventions related to the digital economy. This is 31 thousand more compared to the result recorded a year earlier.
In 2023, the number of patents for inventions registered in the main sectors of the Chinese digital economy reached 406 thousand, which is 45% of the total number of patents issued in the country. At the same time, the average annual growth rate for the five-year period was approximately 21%. Foreign businesses are also expanding their patent portfolios in major sectors of China's digital economy. By the end of 2023, a total of 93 countries (regions) in the PRC had valid patents for inventions in these areas.
Hai Fen Wang of Baidu notes that the boom in AI as a whole can be described as a big leap in the creation of functional applications thanks to a technological breakthrough in big data processing, an increase in the computing power of computers and further successes in creating algorithms. According to him, there comes a time when AI-based applications are starting to play a real role in the economy.[6]
China Mobile buys 8,000 AI servers for $2 billion
On April 24, 2024, it became known that the Chinese telecommunications company China Mobile, serving more than 1 billion customers, will purchase about 8 thousand productive servers for tasks related to artificial intelligence. The total cost of equipment may exceed 15 billion yuan ($2.07 billion at the exchange rate as of the specified date). Read more here.
Mandatory registration and security check. China imposes 24 rules for AI developers
On August 15, 2023, authorities China published a set of rules to regulate the booming industry. artificial intelligence The document contains 24 items that cover a variety of areas, including security and mandatory registration.
According to Bloomberg, Beijing is trying to find a balance between strict state control and support for the AI industry in the face of fierce market competition. This is necessary to ensure that Chinese companies remain viable in the global market. Seven agencies, including the Cyberspace Administration of China and the National Development and Reform Commission, will assume responsibility for overseeing the implementation of the requirements.
The approved norms include mandatory labeling of content generated by AI systems: these are images, video materials, etc. This measure is focused on the fight against deepfakes. China will also require any company implementing an AI model to use "legitimate data" to train its models and disclose that information to regulators upon request. In addition, Chinese AI companies must provide a clear mechanism for dealing with public complaints about services or content.
| | AI can fundamentally change the way people work, live and enjoy themselves. But if AI development continues without proper oversight, it will create clear risks and threats to humanity, said Andy Chun, an artificial intelligence expert and adjunct professor at the City University of Hong Kong. | |
The final list of rules is less stringent than the original version presented in April 2023. For example, the authors of the document decided to abandon the imposition of fines on AI companies in the amount of 100 thousand yuan ($13.7 thousand at the exchange rate as of August 16, 2023).[7]
Baidu, ByteDance, Tencent and Alibaba ordered $5 billion in processors for generative AI
On August 10, 2023, it became known that the Chinese companies Baidu, ByteDance, Tencent and Alibaba ordered Nvidia chips for $5 billion to develop and maintain projects in the field of generative artificial intelligence. Read more here.
China has developed a national standard for large language models
In early July 2023, it became known that China had developed a national standard for large language models (LLM). This is expected to help regulate the artificial intelligence industry.
It is reported that the creation of the document is supervised by the Chinese Institute for Standardization of Electronics under the Ministry of Industry and Informatization of the PRC. The project also involves specialists from Baidu, Huawei, 360 Security and Alibaba. A specially formed government structure will be responsible for the implementation of the standard in the field of LLM. Thus, Beijing seeks to minimize potential violations in the field of AI and stop the uncontrolled development of the market.
Details about the document were not disclosed. But it is noted that the Chinese authorities intend to stimulate the development of promising solutions based on LLM, since they can "contribute to economic growth and become a useful everyday tool." At the same time, it is said about the need to minimize the risks associated with the active introduction of AI in various areas. Among the industries for which AI can play the most important role, China includes office work, biopharmaceuticals, remote sensing of the Earth and meteorology.
After the startup, OpenAI backed by, Microsoft launched a chatbot, ChatGPT major Chinese tech firms began actively developing similar solutions. Beijing has stated the need to develop AI to improve labor productivity and stimulate post-pandemic growth. The Chinese Academy of Information and Communication Technologies (CAICT) announced the acceleration of the development of the LLM direction. However, China's Cyberspace Administration (CAC) has yet to license any AI generative product in the country, even though big tech firms like Baidu, Alibaba and iFlytek have introduced ChatGPT counterparts.[8]
China is the first of the countries to introduce regulation of generative artificial intelligence
On July 13, 2023, the Chinese authorities were the first in the world to introduce special rules governing the activities of generative artificial intelligence. It is expected that this initiative, among other things, will help raise security in the relevant area.
The development of the document was attended by the Office of Cyberspace of China, the National Commission for Development and Reform of the PRC, the Ministry of Education of China, the Ministry of Science and Technology of the PRC, the Ministry of Industry and Information Technology of the country, the Ministry of Public Security of China and some other structures.
The rules, as they say, are developed in order to promote healthy development and standardized application of generative AI, protect national security and social public interests, as well as protect the legal rights and interests of citizens, legal entities and other organizations. The measures apply to the use of AI technology to provide services for the generation of text, images, audio materials, video and other content for the population in China.
| | The state adheres to the principles of giving equal importance to development and security, promoting innovation and legal management, takes effective measures to introduce generative AI, carries out inclusive discretion and classified and hierarchical supervision of these services, the document says. | |
The provision and use of AI generative services must be carried out in accordance with laws and administrative provisions, respect for public morality and ethics. Services in the relevant area will have to obtain a license to start their systems. In addition, it is necessary to respect intellectual property rights, business ethics, keep trade secrets and not use algorithms, data, platforms and other tools to carry out unfair competition.[9]
The founder of the manufacturer of household appliances Midea created a scientific fund for $428 million for AI projects
On May 21, 2023, Chinese billionaire He Xiangjian, founder and largest shareholder of home appliance manufacturer Midea Group, announced the creation of a scientific fund to develop projects in the field of artificial intelligence and ecology. The financial support of the new structure at the initial stage is 3 billion yuan (approximately $425.8 million at the exchange rate as of May 23, 2023). Read more here.
Chinese counterparts of ChatGPT are created with tough political censorship
In early May 2023, it became known that the Chinese authorities oblige companies providing access to chat bots based on artificial intelligence to comply with strict censorship requirements. Read more here.
China has released a list of requirements for developers of services like ChatGPT
On April 11, 2023, it became known that it China intends to introduce new requirements, according to which all generative-based services artificial intelligence will be tested before the actual launch. This could create uncertainty about the future of bots like those ChatGPT represented by the country's largest technology companies, including Baidu Inc.
According to Bloomberg, service providers in the PRC must ensure that their AI systems provide reliable information and do not violate copyright. In addition, safety requirements must be met. China's State Internet Information Office (CAC) has already drafted guidelines for regulating the AI industry. In particular, operators of AI platforms will have to label content generated by neural networks without fail.
The CAC initiative complements Beijing's attempts to control the explosive growth of the generative AI-based services market. China does not hide its desire to raise the level of development of the AI industry, but at the same time plans to apply control measures. There is a possibility that the PRC will ban foreign artificial intelligence services developed by companies such as OpenAI and Google. At the same time, the Chinese authorities will avoid creating excessive pressure on local corporations, fearing to crush the emerging and very promising market for generative AI.
Experts say that the rules defined in the CAC document will affect how AI models develop in China. In the future, this may lead to the introduction of censorship tools in the relevant area. In general, as noted, Beijing plans to introduce rules governing the use of neural networks in a number of key industries.[10]
Chinese army begins simulating invasion of Taiwan with artificial intelligence
On January 15, 2023, it became known that the Chinese army was starting to model the invasion of Taiwan using artificial intelligence. According to the South China Morning Post, the People's Liberation Army of China (PLA) will conduct a blockade exercise around Taiwan in 2023. For this, it is planned to use information technologies. In particular, the AI will simulate the situation of deterrence of US forces in Taiwan. Read more here.
2022
China overtakes the United States in the number of scientists in the field of AI
At the end of 2022, China accounted for approximately 38% of the leading researchers in the field of artificial intelligence working in American institutions. For comparison, in the United States, this figure is estimated at 37%. In 2019, the values were 27% and 31%, respectively. Such figures are given in a study by MacroPolo, the results of which were published in early March 2024.
The report said that in 2022, approximately 47% of the total base of leading AI professionals were of Chinese origin. Another 18% fell on the United States, approximately 12% - on European states. At the same time, it was in the United States in 2022 that the most leading researchers in the field of AI worked - about 42%. Approximately 28% fell on China, approximately 12% - on European countries.
If we consider the segment of the most elite scientists in the field of AI, then in 2022 about 28% of them were of American origin. For comparison, in 2019 this figure was 35%. The PRC in 2022 accounted for 26% of the most outstanding AI researchers, while three years earlier the value was only 10%. Such dynamics reflect the efforts of the Chinese authorities to develop artificial intelligence technologies in the country.
AI platforms are rapidly evolving around the world, the study says. At the same time, the main difficulty is to balance the capabilities of AI to solve a number of economic and social problems with the shortcomings that cover the relevant tools. The fact is that AI tools can be used to create deepfakes, violate information security and automatically generate malicious software code. Companies and organizations around the world are striving to monetize the potential of AI, and research in this area is gaining momentum.[11]
China outpaces US in AI patents
China is increasingly ahead of the United States in the number of AI patents, which indicates the Asian country's determination to shape and influence technology that could have wide-ranging consequences for the world's richest economies.
China introduced mandatory digital labeling of texts, photos and voices created by artificial intelligence
As it became known in December 2022 China , mandatory digital labeling of texts, photographs and voices created has been introduced. artificial intelligence According to the Office of the Central Commission for Internet Affairs of China, AI can be used to disseminate illegal information, slander and fake personalities, and labeling will avoid this.
Under the rules, new "deep synthesis" digital products will be subject to a safety assessment by the government. Each product must be deemed compliant before being released. The administration specifically requires the presence of obvious "tags," for example, watermarks that denote content created by AI.
| | Services providing functions such as intellectual dialogue, synthesized human voice, human face generation and immersive realistic scenes that generate or significantly change information content should be prominently marked to avoid confusion or incorrect identification by the public, the Office of the Central Commission for Internet Affairs of China said in a statement. | |
Companies that provide deep synthesis technology must maintain a register of such products in accordance with the law, and people using the technology must register accounts under their real names so that their actions on the use of synthesized data and images can be tracked.
In addition to restrictions on deep synthesis services for the production and dissemination of information, the new rules also indicate the need to implement a real identity authentication system, as well as other control systems, such as user registration, algorithm mechanism verification, data security. In addition, technical safety measures must be established.[12]
2021: Chinese artificial intelligence has become an obstacle to the work of CIA agents in the country
According to current and former officials, published by November 10, 2021, the rapid development of technologies in the PRC, including artificial intelligence, is increasingly impeding the work of CIA agents in the country.
They say an China increasingly complex target, more opaque, while demand for understanding decision-making Xi Jinping is growing and tensions USA with are heating up over many issues from Taiwan to high technology.
Among other obstacles posed by the Chinese special services to American intelligence, other "greatly increased (technical) capabilities" of counterintelligence of the PRC, as well as a lack of employees who speak Chinese, are called. All this makes it difficult not only to obtain data on China's strategy towards Taiwan or Beijing's defense spending, but even information on who can be considered the successor to current Chinese leader Xi Jinping.
Former officials explained that restoring the CIA network in China is a long process that involves recruiting and implementing new assets, followed by a systematic expansion of assets' access to sensitive information.
In addition, CIA officers in China face significant challenges posed by a growing surveillance system that has covered Chinese cities with more cameras and uses sophisticated software and AI to recognize faces and track threats.
In a July 2021 interview with National Public Radio, Central Intelligence Agency director Bill Burns said the agency was looking into how to deal with "widespread technical surveillance" and other "very advanced Chinese intelligence capabilities."
The Biden administration is seeking to shift its foreign policy strategy more toward countering China, Burns in October 2021 announced the creation of the China Mission Center to increase the agency's focus on an "increasingly hostile Chinese government."
Chinese Foreign Ministry spokesman Zhao Lijian called the creation of the CIA Mission Center "a typical sign of the Cold War mentality." The US "should view China's development and Sino-US relations in an objective and rational light and stop doing things that damage mutual trust and cooperation," he added.
According to officials, a stronger turn towards China will not happen soon. Pentagon China's nuclear arsenal is now growing faster than forecast, the latest in a series of heightened assessments of Beijing's global ambitions, it said last week.[13]
2020
China bans free export of AI technologies
At the end of August 2020, the Chinese Ministry of Commerce banned the free export of AI technologies, including text analysis technologies, content personalization and speech recognition. The new restrictions could delay the conclusion of a deal to sell TikTok by the Chinese company ByteDance abroad.
The Commerce Department has released an updated export control list, which now includes "personalized content recommendations based on data analysis" and a range of other technologies. It is understood that they can be used for both civilian and military purposes, so their sale abroad should be limited.
TikTok's video streaming app, which offers users content based on their preferences, is not mentioned in the list presented, but its algorithms use some of the listed technologies. The list of restrictions has not been revised for more than 10 years, so the media associate the changes with the conflict between the United States and China over the sale of TikTok, which interested companies such as Microsoft, Oracle and Walmart. When the updated list will take effect is still unknown.
| | The new restrictions may mean that the Chinese government wants to take a wait-and-see attitude and assess what actions the US and other governments can take given the growing regulatory restrictions in the hardware and software industry, said Ross Darrell Feingold, a lawyer and political risk analyst who advises multinationals. - These proactive restrictions, which actually act as a ban on the export of certain AI technologies, will help Chinese companies in the future avoid a situation like the one around the sale of TikTok.[14] | |
Gostelecanal in China began to use a virtual news anchor who is not distinguishable from humans
At the end of May 2020, the Chinese state news agency Xinhua began to use a virtual news anchor on its TV channel, which is not distinguishable from humans. The AI reporter is named Xin Xiaowei, modeled after Zhao Wanwei, who is one of the leading state television channels.
2018: 52.8% rise to 33.9 billion yuan
In 2018, sales of software and equipment intended for the creation and operation of artificial intelligence systems in the Chinese market reached 33.9 billion yuan ($4.93 billion at the exchange rate as of July 16, 2019), which is 52.8% more than a year earlier. Such data were published in July 2019 by the non-governmental organization Internet Society of China.
China accounted for 12.56% of the global AI technology market at the end of 2018, while in 2017 this share was measured at 9.41%.
By the end of 2018, there were 3,341 artificial intelligence companies in China, more than one-fifth of the world's total (15,916 ), Xinhua News Agency said.
The Chinese Internet Society report also said spending on cloud computing, big data and the Internet of Things in 2018 was 90.7 billion yuan, 540.5 billion yuan and 1.2 trillion yuan in China, respectively.
According to IDC estimates, computer vision in 2018 became the largest segment of the Chinese artificial intelligence technology market with revenue of $750 million. The top three categories also included developments in the field of security and video surveillance and facial recognition system for the financial sector.
Chinese President Xi Jinping is taking measures aimed at deeply integrating big data and artificial intelligence into the real economy to stimulate the country's long-term growth.
The development of AI plays a key role in the implementation of the PRC state plan "Made in China-2025," which involves the modernization of production to achieve a high level of automation and manufacturability. China has made the development of artificial intelligence a national priority, aiming to become a leader in this area by 2030.[15]
Notes
- ↑ Registered users of large generative AI models in China exceed 600 million, data shows
- ↑ AI ‘coaching’ replacing human instructors at hi-tech Chinese driving school
- ↑ US lens on investments to hinder AI, chip sectors
- ↑ Tech war: US wants cloud firms to report foreign users tapping their computing power for AI applications
- ↑ China home to over 4,500 AI companies: vice minister
- ↑ China owns 378,000 AI patents, rising faster than global average
- ↑ Beijing Tries to Regulate China’s AI Sector Without Crushing It
- ↑ China to create and implement national standard for large language models in move to regulate AI, while using its power to transform industries
- ↑ Interim Measures for the Management of Generative Artificial Intelligence Services
- ↑ China Mandates Security Reviews for AI Services Like ChatGPT
- ↑ The Global AI Talent Tracker 2.0
- ↑ The State Internet Information Office and other three departments issued the "Regulations on the Administration of Deep Synthesis of Internet Information
- ↑ China Is Evading U.S. Spies — and the White House Is Worried
- ↑ China's new AI export curbs threaten TikTok's US sale
- ↑ China’s AI market to hit 71 bn yuan
| The site content is translated by machine translation software powered by PROMT. The machine-translated articles are not always perfect and may contain errors in vocabulary, syntax or grammar. Read original article If you find inaccuracies or errors in the results of machine translation, please write to editor@tadviser.ru. We will make every effort to correct them as soon as possible. |