China and the United States have not yet ended the large -language model competition around ChatGPT. The American artificial intelligence giant Openai, which was born on February 16, Sora, and the second track for Sino -US videos to expand Wensheng video.

SORA can generate a real animation video with the fake real animation videos according to the text instructions.This marks that OPENAI has further made a major breakthrough in the field of multi -mode and large language model technology on the basis of the large language model of ChatGPT.

Although the concept of Wensheng video and virtual videos itself is not new, Sora's video duration has increased to 60 seconds, which is nearly 10 times that of other existing technologies.Sora's ability to make other players can't reach the dust, including complex multi -lens switching, reflexes and light that conforms to physics, can create delicate and emotional scenes, characters, characters, and natural smooth dynamics; and in the past, in the pastThe "one -glance fake" AI video formed a strong contrast, and there is huge potential to subvert film and television, short videos, advertising and other industries.

The results of

Sora's leaps have caused some Chinese scientific and technological operators to worry about the extension of the level of artificial intelligence in China and the United States, and also promote the acceleration of Chinese players in the field of video generation to seek breakthroughs.

The State -owned Assets Supervision and Administration Commission of the State Council of the State Council of the State Council held less than a week after SORA's emergence, and held a central enterprise's artificial intelligence special promotion meeting, requiring central enterprises to accelerate the layout and development of the intelligent industry, and accelerate the construction of a number of intelligent computing power centers.

Alibaba Intelligent Computing Research Institute released a new generation AI model EMO on February 28. You can use a portrait portrait photo and audio to let the characters in the photo sing and speak according to the audio content.

Baidu has released a video generation model called Univg in January this year.Li Yanhong, the founder, chairman and chief executive officer of Baidu, also on February 28 at the fourth quarter and annual performance meeting of Baidu 2023. The integration of multi -modal or multi -mode such as video to videos is a future basic model.Very important development direction.

Li Yanhong emphasized that this is a necessary direction for GM artificial intelligence. Baidu has invested in these fields and will continue to invest in the future.

360 Group founder Zhou Hongyi also shouted to market confidence in an interview with the surging news on February 23 that the gap between China and the United States on AI is mainly reflected in the direction. Once the direction is correct, Chinese companies can catch up in a year or two in one or two years.superior.

Global Market Insight Company Report shows that enterprises and institutions will use Wensheng video to make marketing content more than ten years.In 2022, the global Wensheng video market size is estimated to be US $ 122.5 billion (S $ 165 million); this market is expected to reach more than 35%of the compound annual growth rate from 2023 to 2032, and it is expected to reach $ 2 billion by 2032.

Artificial intelligence has now become one of the most important battlefields in the Sino -US scientific and technological warfare. Whether China can cultivate video generation technology that can catch up with SORA is still unknown in occupying a place in the huge Wensheng video market; in this game; in this game;In the chase, China must also remove two major obstacles.

In October last year, the US Department of Commerce refers to the field of artificial intelligence in China, restricting Nvidia's sale of high -end AI chips to China, including two chips A800 and H800 customized for the Chinese market.

Although Nvidia launched three new chips a month later as alternatives of A800 and H800, the momentum of the United States' constraints from the development of China's AI development from the supply chain is difficult to reverse; Chinese AI companies face huge supply chain risks and must be in the risk of supply chain. They must be in the risk of supply chain.While focusing on R & D technology, it is likely that the scientific and technological enclosure of the West is constantly looking for ways, so it is likely to slow down the progress of technological development.

On the other hand, the world is currently facing the serious lack of AI -related regulations; after the release of Sora, the problems of deep fakes, fake information, and fraud caused by video generation technology have become increasing challenges.

Within a week after the release of

Sora, China happened to appear in China, Olga Loiek, which was stolen on the Chinese network to detonate Chinese netizens' discussion on AI security.The illegal people use the deep pseudo technology to create an Internet celebrity blogger with a face of Live, saying that this deep -pseudo character is Russians who settle in China, sharing various short videos about Russia and Chinese life and bring goods.Obtain traffic.

China is a global leader in AI supervision. When AI technology is still immature, China has released documents to promote the new generation of artificial intelligence development planning to promote the development of big data development.EssenceHowever, the Royk incident reflected that there were still many loopholes in regulatory measures.

China is going to catch up with the United States in the AI ​​competition. It is not impossible. It is just facing the increasingly steeper uphill road, not only facing the competition in the United States, but also that it is difficult to escape the AI ​​governance problems faced by countries.Sora is a node that has further expanded the gap between China and the United States after ChatGPT, and it is also an important reminder -this AI competition does not have a midfielder. China can only turn over through unremitting catch -up, otherwise, if you don't pay attention, you will not pay attention, and you will not pay attention, and you will not pay attention.It will be left behind.