LLM is the deep learning AI algorithm behind chatbots like ChatGPT that can recognize, summarize, translate, predict, and generate content using very large data sets.
Export restrictions on semiconductor chips used in LLM processing are one of the major obstacles to China's generative AI development.
Citing national security concerns, the U.S. will ban companies including Nvidia and AMD from selling their most powerful chips and the equipment used to make them to China in 2022, forcing them to sell their most powerful chips and the equipment used to make them to other countries. urges them to follow suit.
Zeng Yi, a professor at the Chinese Academy of Sciences and head of China Electronics Corporation, said that as a result of this push by the U.S. government, Chinese companies are focusing not only on the advancement of LLMs but also on how to develop new hardware and software. At the panel discussion, he said that it was necessary to
“I believe that artificial intelligence is still the key to trends in the world industrial chain. However, we must prepare to not rely on these technologies for hardware infrastructure that may be disconnected from us. We also need to do that,” Zeng said.
The United States and China are locked in a fierce competition for supremacy in the field of AI, with China's lead expected to expand in areas such as facial recognition and autonomous driving.
But generative AI has been transformed since OpenAI introduced ChatGPT in November 2022, the first ever human-like chatbot trained on large datasets covering topics from history to computer code. has been achieved.
Chinese scientists say AI model can mark optimal locations for solar panels
Chinese scientists say AI model can mark optimal locations for solar panels
Christopher Thomas, a senior fellow at the Beijing-based Brookings Institution, echoed this on a separate panel on AI governance on Thursday, saying: The United States ranked first, with China a distant second. ”
“In a world where the United States does everything its own and China does everything its own, we're going to end up wasting money on duplication instead of solving problems,” he said.
But Mr. Yuan was optimistic about a breakthrough. He said that no single country or company can completely and independently “own” artificial intelligence, because each time a new product is introduced, competitors develop similar or more advanced chips. Ta.
In addition to the U.S. chip ban, China also faces domestic challenges such as huge costs of computing and censorship.
Zeng said the risk-averse approach was the biggest challenge for China's AI development, and questioned whether people were willing to challenge the majority view.
“I think the biggest challenge with artificial intelligence in China right now is that we may not be good at taking bigger risks. We are not going to pay the cost of … This is the biggest core of artificial intelligence development in China.”