Author of this article: Zhang Yifan
Editor: Shen Siqi
Source: Hard AI
From July 4 to July 6, 2024, the three-day World Artificial Intelligence Conference (WAIC) will be held in Shanghai. The continuous hot weather did not stop everyone's enthusiasm. Despite the rising temperature, the venue was still full of people.
Experts from academia and industry conducted in-depth discussions on key issues such as the development prospects, challenges, and governance of AI. From Xue Lan and Zhou Bowen of Tsinghua University, to Synopsys CEO Gai Sixin, to Blackstone Group founder Su Shimin, and three Turing Award winners Roger Reddy, Manuel Blum, and Yao Qizhi, the participating experts examined the importance of AI from different perspectives.
GaiSIN — President and CEO of Synopsys
In the next 8-9 years, AI will double the growth of the semiconductor industry
Insights were shared from the perspective of the semiconductor industry. He pointed out that AI is driving rapid growth in the semiconductor industry, and sales are expected to double from 500 billion US dollars to 1 trillion US dollars within 8-9 years. Gaisnew also emphasized the importance of considering safety, compliance, and energy efficiency in the chip design process.
Su Shimin — Chairman, CEO and Co-founder of Blackstone Group
AI changes the way companies are valued, and the impact of AI on investment targets should be assessed from a 5-10 year cycle
Blackstone Group founder Su Shimin shared the impact of AI on the investment sector. He believes that AI is changing the way companies are valued. Some companies are benefiting from AI integration, while others are facing challenges.
Su Shimin emphasized the importance of considering the long-term impact of AI in investment decisions, and suggested evaluating the potential impact of AI on investment targets on a 5-10 year cycle. He pointed out that investors need to be wary of the disruptive changes AI may bring, while also being aware of new opportunities created by AI. Su Shimin advised investors to pay attention to companies that can effectively integrate AI, while being wary of industries that may be eliminated due to AI development.
But in addition to growth, participants also hinted at the risks behind AI.
Three Turing Award winners — Roger Reddy, Manuel Bloom, and Yao Qizhi
Roger Reddy: AI could lead to multi-fold growth in productivity and GDP
Roger Reddy called for a balanced view of the development of AI. He pointed out that every new technology brings opportunities and challenges, and development should not be abandoned because of potential risks. He predicted that AI could lead to multi-fold increases in productivity and GDP. It emphasized the importance of AI as a tool and called for more attention to how AI can be used to enhance human capabilities. Reddy emphasized the need to invest in research to explore how to make better use of AI technology, make AI a tool to enhance human intelligence, and help people work 10-100 times more efficiently.
Manuel Bloom: The CTM model inspired AI development
Manuel Bloom shared his experience in the field of consciousness research over 20 years. He introduced the “Conscious Turing Machine” (CTM) model he developed with his colleagues. This model likens the brain to a theater where viewers (neurons) listen to what's happening on stage (conscious thoughts). There are people on stage speaking, asking questions, and answering; this is a broadcast to all viewers. When consciousness occurred, all the listeners heard the conscious ideas spread. Bloom emphasized that the CTM model is unique in that it does not have a central decision maker, but rather allows every “participant” to contribute important information. This model provides a new perspective for understanding human consciousness and provides inspiration for the development of general artificial intelligence.
Yao Qizhi: Focus on the risks posed by AI and study how to coexist with it
Yao Qizhi focused on the potential risks posed by AI and its governance issues. He classified AI risks into three categories:
1) Expanding cybersecurity risks
2) Unknown risks brought about by rapid changes in social structures
3) Surviving or being at risk
Yao Qizhi pointed out that as a computer scientist, he is particularly concerned about the first and third types of risks. He stressed that with the advent of AI, the challenges facing cybersecurity are similar to those of 40-50 years ago and require expert attention. At the same time, he also expressed concern about the risks that AI may bring. He believes that humans have created a new species that may be much more powerful than themselves, and how to coexist with it is a major challenge.
Xue Lan —— Dean of Su Shimin College, Tsinghua University; Dean of the Institute of Artificial Intelligence and International Governance, Tsinghua University
Positive and negative, the dual nature of artificial intelligence
Using the 169 specific indicators of the United Nations Sustainable Development Goals (SDG) as an example, Xue Lan pointed out that artificial intelligence may have a positive effect on 134 (79%) of them, but it may also adversely affect 59 (about 35%). These adverse effects include intrinsic technology issues, data security, algorithm discrimination, etc., all of which need to be controlled through effective governance.
Zhou Bowen — Director of the Shanghai Artificial Intelligence Laboratory and Chair Professor Hui Yan at Tsinghua University
Balance AI safety and performance
The concept of “AI 45 degree balance ratio” was proposed. He emphasized the need to find a balance between the safety and performance of AI, and proposed the “causal ladder of trusted AGI” technical path, which includes three stages of broad alignment, interventionability, and reflection.