The Shanghai Securities News - China Securities Network (Reporter Yu Xiangming) Liu Liehong, a member of the Party Group of the National Development and Reform Commission and director of the National Data Administration, said at a series of theme press conferences on high-quality completion of the 14th Five-Year Plan held by the Information Office of The State Council on August 14 that the application scale of artificial intelligence in China is growing rapidly. At the beginning of 2024, the average daily consumption of tokens in China will be 100 billion. As of the end of June this year, the average daily consumption of tokens has exceeded 30 trillion, increasing by more than 300 times in one and a half years. The so-called Token, which is commonly referred to as a word element, serves as the smallest data unit for processing text, much like "traffic" in the Internet era.

Liu Liehong pointed out that data plays a crucial role in promoting the "Artificial Intelligence +" process, especially the construction of high-quality datasets is of vital importance. For instance, in the medical and health sector, by using high-quality labeled medical image datasets, the disease diagnosis accuracy of models can be increased by more than 15%.

He said that our country is continuously promoting the construction of high-quality datasets. As of the end of June this year, China has built over 35,000 high-quality datasets, with a total volume exceeding 400PB (1PB can store approximately 500 million high-definition photos of 2MB in size, and the total volume of 400PB is about 140 times the total digital resources of the National Library of China).

It is reported that the training of artificial intelligence models has also driven up the demand for data trading in China. As of the end of June this year, the cumulative transaction volume of high-quality datasets across the country has reached nearly 4 billion yuan, and the total scale of high-quality datasets listed by data trading institutions has reached 246PB.

Take the Beijing Data Exchange as an example. The proportion of high-quality datasets in the total trading volume has jumped from 10% last year to nearly 80% at present. Shanghai, Tianjin, Anhui and other places are piloting new models such as "valuing data corpora as equity", guiding enterprises to convert high-quality data sets into equity and invest them in relevant enterprises. The development of high-quality datasets requires the support of the data annotation industry. China has established seven data annotation bases in cities such as Chengdu, Shenyang, and Hefei to facilitate the construction of high-quality datasets.

Liu Liehong stated that in the next step, they will continue to promote the construction of high-quality datasets through systematic layout and accelerate the creation of data highlands in key areas such as embodied intelligence, low-altitude economy, and biomanumanufacturing.

https://finance.eastmoney.com/a/202508143484067120.html