TMTPOST—The emergence of AI has made managing data security 100 times more difficult, said Andrew Chi Chih Yao, a Turing Award winner and a professor of computation at Tsinghua University.

Yao, who is also a member of the China Academy of Sciences, made the comments during the 2024 World Artificial Intelligence Conference (WAIC) and the High-Level Meeting on Global Governance of Artificial Intelligence held from July 4 to July 6 in Shanghai.

Addressing data security-related risks requires the involvement of experts, scientists, as well as governments, lawyers, economists, and people from every industry, he added, quoting A.M. Turing as saying, “these risks are unpredictable, and we cannot predict what machines will do once they have enough computing power.”

In Yao's view, AI amplifies Internet-related risks, and its risks are also related to computing power. Current computing power can solve many problems that were previously unsolvable, but this is also what makes it more frightening.

Regarding AI perspectives and concerns, another Turing Award winner Raj Reddy noted that everyone is concerned about AI governance risks and safety issues, but there is still no clear answer as to what research and investments are needed.

Reddy believes that new technologies will bring new opportunities and problems, but human beings should not stop progressing due to fear of negative impacts. In the future, if AI can increase everyone's work efficiency tenfold, the global GDP could rise from $100 trillion to $1,000 trillion, he noted.

Zhou Bowen, the new director of the Shanghai AI Laboratory, highlighted the rapid development of generative AI, represented by large models, and the accompanying risks such as data leaks, ethical issues, and the potential for AI to become uncontrollable.

"Currently, our investment in AI safety is far behind our investment in AI performance. Only 1% of the world's resources are invested in alignment or safety considerations," Zhou said. He emphasized the necessity for AI to ensure controllability and development alongside safety, advocating for "TrustWorthy AGI (General Artificial Intelligence)."

Robin Li, the founder and CEO of Baidu, addressed the state of AI development, critiquing the industry's focus on basic models and rankings over practical applications. "With only a foundation model, whether it is open source or closed source, is worthless Without application," he said, urging companies to prioritize applications over model development.

Li highlighted the success of Baidu's Wenxin large model, which has seen significant usage growth. He also noted the debate between open source and closed source models, arguing that closed source models offer superior capabilities and are more suitable for competitive market environments.

Wang Jian, an academician of the Chinese Academy of Engineering, spoke at the WAIC 2024 roundtable about the untapped potential of GPT. He reminisced about the long history of AI development and expressed optimism about the next decade, which he believes will be transformative for AI.

Wang emphasized that new technologies inevitably lead to the emergence of large companies and the rebirth of existing ones. He urged large enterprises to view AI as a revolutionary tool, which will drive significant industry changes.

Jing Xiandong, Chairman and CEO of Ant Group, spoke about the impact of AI on industries, likening it to the transformative effect of the internet. He highlighted the role of professional intelligent agents in overcoming the limitations of general large models in rigorous industries, promoting service upgrades, and building a professional intelligent agent ecosystem.

Jing emphasized the importance of making AI as convenient as scanning code payments, aiming to extend the benefits of AI technology to more people.

Xiao Song, Global Executive Vice President of Siemens, discussed the impact of AI on the industrial sector. He highlighted Siemens' collaboration with Microsoft to develop generative AI for industrial scenarios, which has significantly lowered programming thresholds and shortened development times.

Xiao showcased the successful implementation of AI projects in Siemens' Chengdu digital factory, leading to significant improvements in manufacturing costs and process quality.

Xu Li, Chairman and CEO of SenseTime Technology, pointed out that the AI industry has not yet reached a "super moment" due to the lack of vertical industry applications. He emphasized that applications are crucial for driving the next phase of AI development and highlighted the importance of synthetic artificial data and natural, delay-free interaction for enhancing AI models.

Xu called for a collaborative effort to promote the AI super moment through practical applications and continuous improvement of AI models.

Liu Qingfeng, Chairman of iFlytech, discussed the impact of AGI on education. He highlighted global efforts to integrate AI into education, such as the release of guidelines by UNESCO and regulatory provisions by the White House.

Liu showcased the progress of the iFlytech Spark V4.0 model, which has significantly enhanced educational tools and improved learning outcomes. He emphasized the need for education to adapt to the AI era and cultivate new talent capable of leveraging AI technologies.

Meng Pu, Chairman of Qualcomm China, discussed the synergy between 5G and AI, predicting that transferring some generative AI workloads to the terminal side could save significant computing resource costs. He stressed the importance of high-performance AI processors and optimizing AI models for terminal devices to promote widespread generative AI applications.

Meng highlighted the potential market growth for AI-integrated devices, forecasting a significant increase in the shipment of AI mobile phones and the penetration rate of AI PCs in the coming years.