Artificial intelligence is a new frontier in human development, and its rapid development has had a profound impact on economic and social progress and the advancement of human civilization, as well as bringing unknown risks and complex challenges.
At the opening ceremony of the 2024 World Artificial Intelligence Conference, Xue Lan, Dean of the Schwarzman College at Tsinghua University and Dean of the Institute for AI International Governance at Tsinghua University, Zhou Bo-Wen, Chief Scientist of the Shanghai AI Laboratory and Huiyan Chair Professor at Tsinghua University, Ge Xin, President and CEO of Synopsys, Stephen A. Schwarzman, Chairman, CEO, and Co-Founder of Blackstone Group, Joshua J. Ramo, Chairman and CEO of Sonoma Advisors and Investment Company, Shen Xiangyang, former Executive Vice President of Microsoft and Foreign Member of the National Academy of Engineering, and three Turing Award winners, Andrew Yao, Raj Reddy, and Manuel Blum, conducted in-depth discussions and exchanges around the conference theme of "Promoting Shared Development through Consultation and Good Governance through Good Intelligence."
Xue Lan, Dean of the Schwarzman College at Tsinghua University and Dean of the Institute for AI International Governance at Tsinghua University, stated that research has shown that AI will have a promoting effect on 134 specific targets (79%) and a hindering effect on 59 (35%). Artificial intelligence may bring three major risks: first, the risks inherent in artificial intelligence itself; second, the risks based on the development of technology; and third, the risks of misuse based on the application of technology.
In the past few years, China has established a relatively complete system. At the foundation level, there is a series of laws and regulations, and on the other hand, there are a series of governance rules for algorithms, computing power, and data. A multi-dimensional, multi-level, multi-field, and multi-measure overall governance system has been constructed. Xue Lan emphasized that to promote the development of artificial intelligence and the prevention and control of risks, it is essential for society to have a better understanding of artificial intelligence.
From a global perspective, there are still many challenges in the development of artificial intelligence. The first is the gap in infrastructure construction, the second is the gap in civic literacy, and the third is the gap between the development and governance of artificial intelligence.
Xue Lan called on international organizations such as the United Nations to truly play a comprehensive and coordinated role, to break down various barriers and strengthen international cooperation through joint efforts, while the industry and academia work together to promote the rational development of artificial intelligence.Director and Chief Scientist of the Shanghai Artificial Intelligence Laboratory, and Huiyan Chair Professor at Tsinghua University, Zhou Bo-wen
The rapid development of artificial intelligence, with the continuous improvement of AI capabilities, has brought a series of potential risks and concerns about the models themselves and their applications.
From the public's attention to the order of AI risks, the first is the risk of data leakage, misuse of privacy, and copyright-related content risks; followed by the use risks brought by malicious use, such as forgery and false information, which also induce related ethical issues such as bias and discrimination; there are also concerns about whether it will bring systemic challenges to the employment structure and other social systems, and even in science fiction movies, there are settings of AI out of control and humans losing autonomy.
Zhou Bo-wen said that some of these AI risks have already emerged, but more are to prevent these risks in a potential way, which requires the joint efforts of science and technology, and requires the scientific community to make more contributions.
Overall, the investment in AI model security is far behind AI capabilities. Currently, 99% of the world's computing power is used for pre-training models, and less than 1% is used for security-first considerations.
Zhou Bo-wen proposed the technical thought system of "AI 45-degree balance rate". The 45-degree balance law means that from a long-term perspective, it is necessary to roughly develop along the 45-degree balance between safety and performance. The so-called balance means that there can be fluctuations in the short term, but it cannot be below 45 degrees for a long time.
The technical thought system of the 45-degree balance requires strong technical drive, full process optimization, multi-subject participation, and agile governance to achieve the AI 45-degree balance rate.
President and Chief Executive Officer of Synopsys, Inc., Aart de GeusAI starts from semiconductors, from chips, and it is driving very remarkable disruptive technologies, while the software above is also running. Synopsys has applications for specific markets, and users experience AI through these applications.
The semiconductor industry took 60 years to achieve sales of 500 billion, and doubled to 1 trillion in 8-9 years, with almost all the incremental growth driven by AI. AI requires a lot of infrastructure to train models. In the past few decades, many disruptive technologies have emerged, such as the internet, personal computers, and mobile devices.
To ensure the safety of AI applications in real-life scenarios, semiconductors need to ensure the security of chips, that is, every IP or part of the material inserted needs to have security guarantees; and also need to improve the energy efficiency of chips.
Synopsys has begun to try to apply AI to the workflow, established an AI Center of Excellence, trained employees to understand the ethics, risks, and values of AI, and used models to develop new chips, and continuously monitor and track.
Gai Xin said that Synopsys has included generative AI content in the design of semiconductors. Since 2017, Synopsys has invested in AI, and the products sold to customers have always included AI. And Synopsys has begun to use AI to help improve the work efficiency of engineers and guide chip design.
Gai Xin said that the pace of AI innovation must be carefully balanced with the need for its responsible development, and the best way to achieve this balance is through global collaboration. Global collaboration can enable people around the world to benefit from the tremendous progress brought about by AI, while ensuring that these innovative capabilities are practiced responsibly.
Artificial intelligence is constantly creating business opportunities, profoundly changing the investment landscape, while simultaneously generating governance issues and governance opportunities. On this issue, Stephen A. Schwarzman, Chairman, CEO, and Co-Founder of Blackstone Group, and Joshua Ramo, Chairman and CEO of Thorney Investment Consulting and Investment Company, discussed.
AI has had some impact on the valuation of existing enterprises. Currently, some companies are doing well in the integration of AI, but some companies are not very adaptable to AI. Schwarzman believes that technological development will have a profound impact, and companies cannot bury their heads in the sand like ostriches, not caring about the risks of technology development outside.
Investors must focus on assets from a long-term perspective of 5-10 years. The current AI technology will change some companies, and no one wants to be a loser after the development of AI technology. Schwarzman said that AI has become an important way for investors to view things.
An interesting point is that in terms of AI, universities are more disadvantaged in breakthrough progress. Because companies can afford and afford the development of the required algorithms, so although there are many academic big shots in universities conducting research, this competition is at a disadvantage compared to the business world.Former Microsoft Executive Vice President and Foreign Associate of the National Academy of Engineering in the United States, Harry Shum, engaged in a dialogue with three Turing Award winners, Andrew Yao, Raj Reddy, and Manuel Blum. Through a highly speculative conversation, they delved into the core issues in the field of artificial intelligence that are the focus of global attention, such as the "double-edged sword" nature of artificial intelligence, its interpretability and predictability, and its rigorous foundation and transformative temperament. In response to the conference motion to strengthen global governance of artificial intelligence, the three Turing Award winners expressed a unanimous high degree of recognition, and at the same time pointed out the important value of talent cultivation in dealing with the future risks of artificial intelligence.
Comment