At the 2024 China Computer Conference, experts have opined that AI serves as an assistive tool rather than a replacement for human software developers.
On October 24, the 2024 China Computer Conference (CNCC2024) kicked off in Hengdian, Dongyang, Zhejiang Province, drawing over 10,000 participants from academia, education, and industry within the computing sector. The conference centered around the theme “Developing New Quality Productive Forces, Computing Leading the Future,” featuring over 800 renowned experts, scholars, and technology elites from around the world, including Turing Award winners and numerous academicians.
In his opening remarks, conference chair and Chinese Academy of Engineering academician Chen Chun emphasized, “We are in an unprecedented era of transformation, with the rapid development of information technology driving a wave of industrial revolution.” He pointed out that computing, as a crucial force propelling societal advancement, is intricately woven into our lives like never before, ushering in a new era of ubiquitous computing characterized by the convergence of the internet of everything, human-machine interaction, and physical objects. Chen suggested that computing not only fosters intelligence but also shapes consensus, offering a new paradigm for scientific research and giving rise to innovative economic models.
Chen stated that high-quality development hinges on the support of new quality productive forces, with computing being the heart and soul of this new productivity. He expressed hope that industry leaders would collaboratively explore new technologies and trends in the computing field, contributing to the cultivation and development of these productive forces.
Professor Mei Hong from Peking University and a Chinese Academy of Sciences academician discussed the evolution of software automation and the challenges posed by current AI advancements in software engineering. “The integration of big data and deep learning has exceeded our expectations, yet many challenges persist,” Mei noted. Over the past decade, his team has made strides in deep learning combined with software engineering, achieving original results in AI-assisted code generation and completion. However, he acknowledged that the application of large language models in code generation is still limited, particularly in supporting programming languages’ structures and semantics, as well as facing hurdles in acquiring and using proprietary code data resources.
“I believe the current technical pathways of large language models are unsuitable for handling and generating program code. We need to integrate existing software engineering techniques and knowledge,” Mei explained. He stressed that while big data and deep learning offer powerful resources and analytical methods, relying solely on these is insufficient for achieving software automation. “Software embodies human civilization and must be firmly in human hands; AI should be an assistive tool, not a replacement for human software developers,” he emphasized.
Lenovo Group’s Senior Vice President Rui Yong suggested that effectively implementing artificial intelligence requires a comprehensive consideration of four key elements: data, algorithms, computing, and infrastructure. He advocated for a mixed framework approach to tackle these challenges. He showcased the limitations of AI, particularly large models, highlighting their lack of reasoning and planning capabilities, and called for a more integrated and hybrid approach that combines various schools of thought and technologies to realize the true potential of AI.
This year’s conference spans three days and features 118 invited reports, three main forums, and 138 specialized forums covering 34 technical domains, including AI, cybersecurity, basic software, education, and semiconductors. The China Computer Conference, founded in 2003, serves as an annual gathering for professionals from academia, technology, industry, and education to exchange insights in the field of computing.