China News Service, Beijing, March 25 (Wang Mengyao) Currently, artificial intelligence has become the development focus of various countries. How to coexist with artificial intelligence and better benefit human society while ensuring its safety and controllability has become a common global challenge.
On the 24th, people from the political, academic and business circles gathered at the China Development Forum’s “Artificial Intelligence Development and Governance Seminar” to discuss artificial intelligence security governance.
Zhang Hongjiang, founder and founding chairman of Beijing Zhiyuan Research Institute, used the word "shocking" to describe Sora.
He said that Sora can provide a realistic scene based on text that cannot be shot in the real world. The multi-modal large model can also judge whether the video is artificially generated by "cherry blossoms and snow cannot appear at the same time" and the uniformity of snow flakes, and can even think , reasoning instead of mechanically executing orders in the past.
"The greatest invention made by mankind is artificial intelligence. There is no doubt that it is also the riskiest and most dangerous innovation." Andrew Forrest, executive chairman of Fortescue Metal Group, said that human DNA will not even exist for decades. It will take hundreds of years to make some improvements, but AI can continue to improve at the speed of light.
"AI's learning ability exceeds the average learning ability of humans." Zhang Hongjiang also said that human learning ability is a gentle straight line, but machine learning ability is an exponential curve. When the two lines intersect, human learning ability will lag behind that of machines. . What's even more frightening is that when the system itself has the ability to improve itself, it may bring about global catastrophic consequences, he reminded.
Zheng Yongnian, dean of the Qianhai Institute of International Affairs at the Chinese University of Hong Kong (Shenzhen) and chairman of the Guangzhou Guangdong-Hong Kong-Macao Greater Bay Area Research Institute, made an analogy with the atomic bomb, saying that the atomic bomb is the first tool of self-destruction mastered by mankind, "We are now One more thing” and “it would be even worse if artificial intelligence and nuclear weapons were combined.”
How to avoid such catastrophic consequences and rein in the development of artificial intelligence? This question appears more important than ever.
Joseph Schiffakis, founder of Verimag Labs and winner of the 2007 Turing Award, bluntly stated that some international conferences are discussing the establishment of a global regulatory framework, but the chances of reaching a global artificial intelligence regulatory framework are slim.
However, the recently reached "Beijing International Consensus on AI Safety" has still made a gratifying breakthrough in the supervision of artificial intelligence. Zhang Hongjiang emphasized that the most important thing is to draw a safety red line: artificial intelligence cannot autonomously copy or improve, cannot seek power, cannot assist bad behaviors, especially weapons of mass destruction, biological and chemical weapons, and create large-scale cyber disasters, and cannot deceive. Humans, these red lines cannot be crossed.
Zhang Yaqin, dean of the Intelligent Industry Research Institute of Tsinghua University and academician of the Chinese Academy of Engineering, is more specific. He suggests linking robots to the people or companies behind them in order to trace entity responsibilities, while establishing a regulatory classification system to determine different levels of risk based on risk. regulatory standards.
In continuing to deepen artificial intelligence security governance, the importance of cooperation has been repeatedly mentioned.
Zheng Yongnian said that the United States and China each have comparative advantages in the development of artificial intelligence, and cooperation between the two countries is needed in terms of governance.
Wu Zhaohui, Vice Minister of the Ministry of Science and Technology of China, also pointed out in his speech that all countries in the world should join hands to promote the development of artificial intelligence in a highly responsible manner, adhere to the principle of intelligence for good and intelligence, and ensure that artificial intelligence is always under the control of humans. To achieve safety, reliability and controllability. (over)