Zhou Hongyi: Boost Big Model AI for Industry Growth

Proposed AI Development Strategy by Founder of 360 Group, Zhou Hongyi

In the midst of a new global wave of technological change, large models, as the core engine of artificial intelligence development, are sparking a new industrial revolution. At this crucial juncture, Zhou Hongyi, a member of the National Committee of the Chinese People’s Political Consultative Conference and the founder of 360 Group, proposed during the 2024 Two Sessions to deepen the multi-scenario application of artificial intelligence. He suggested supporting the development of large models in a more vertical and industrial direction to accelerate the formation of new productive forces.

360 Group founder Zhou Hongyi

Zhou Hongyi explained that the competition between China and the United States in the field of artificial intelligence is twofold: a battle against OpenAI’s general large model foundation on one hand, and a differentiation and specialization in the application of large models on the other. Currently, China still needs time to catch up with the United States in general large model core technology. However, in terms of large model applications, 2024 marks the onset of various application scenarios for large models, indicating that China can certainly pave a path of large model development with distinct Chinese characteristics.

Although general large models play a crucial role in the field of artificial intelligence, their application in enterprise-level scenarios often faces challenges such as a lack of industry depth, insufficient integration with business processes, and difficulties in validating “illusionary” content. Vertical large models demonstrate significant advantages in addressing these issues. Zhou Hongyi emphasized that an important direction for China’s large model development should involve leveraging the advantages of industries and scenarios. This would entail integrating large models with business processes and product functionalities, seeking diversified applications across multiple scenarios, and promoting verticality and industrialization.

Regarding the promotion of vertical and industrial application of large models, Zhou Hongyi put forth three specific suggestions:

  • Firstly, scenarios are crucial. Zhou Hongyi recommended that the government and central SOEs (State-Owned Enterprises) lead the way in offering more application scenarios. This would create more opportunities for the development of vertical, small-scale, and cost-effective large models through greater depth and breadth. He also advised enterprises not to rush into using large models but rather progressively transform their businesses with AI. By analyzing specific scenarios and matching the capabilities of large models with business processes, businesses can gradually drive digital transformation, moving from isolated instances to a comprehensive overhaul and incrementally advance towards intelligent business transformation.

  • Secondly, knowledge is essential. The data and knowledge integrated into large models represent only a fraction of human knowledge. Enterprises possess a wealth of “tacit knowledge,” such as strategic planning and product design, which are unique to them. Vertical large models based on this “tacit knowledge” can better address enterprise issues. Zhou Hongyi thus recommended encouraging companies to manage their tacit knowledge before deploying customized AI. By upgrading their enterprise big data platforms into exclusive knowledge platforms, companies can, through vertical training, delve into enterprise-level scenarios and meet their specific needs effectively.

  • Lastly, business integration is key. Emphasizing the importance of business integration, Zhou Hongyi suggested encouraging and guiding enterprises to deeply integrate large models with digital business systems. By aligning them with business processes, companies can fully exploit the value of large models. Zhou Hongyi believes that the integration of large models with business scenarios is pivotal in achieving intelligent transformation. Thus, the government should incentivize enterprises to fuse one or two business scenarios with large models, creating integrative large models closely tied to the business for widespread implementation. This integration would propel the convergence of these large models with digital systems, driving the development of industrial digitization and new industrialization.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.