focus on the two sessions| National People’s Congress deputies actually: It is recommended to accelerate the cultivation of service providers for large-scale models and promote the opening of multiple scenarios and data

Photo source: Visual China

Blue Whale News, March 7 (Reporter Hu Jie)Artificial intelligence technology represented by large models is an important driving force for a new round of scientific and technological revolution and industrial transformation. It is the main front for developing new productive forces and a key focus in building a scientific and technological power.

With the in-depth advancement of large model applications, AI Agents are gradually becoming a new trend in technological evolution and a mainstream paradigm for large model applications. However, due to the insufficient technical maturity of the large model, the weak understanding of the industry by service providers, the insufficient wealth of callable interfaces, and the lack of unified standards for invocation methods, AI Agents at this stage are relatively lacking in autonomy and flexibility. They mainly follow fixed steps to process some tasks with process standards and clear rules. It is still difficult to fully meet the implementation needs in actual business scenarios, and there is still a long way to go from the all-round AI assistant people envision.

“As large models such as DeepSeek make new breakthroughs in reasoning capabilities and open source models, AI Agents are expected to demonstrate stronger self-reflection and independent decision-making capabilities, and can demonstrate greater efficiency in industry applications. rdquo; National People’s Congress representative and chairman of Orient Fortune Information Co., Ltd.(hereinafter referred to as Orient Fortune) actually said.

Blue Whale News reporter learned that in order to better unleash the huge potential of AI Agents to empower thousands of industries and promote the prosperity of the large-scale model application ecosystem, during the two sessions this year, suggestions were actually brought on accelerating the promotion of large-scale model application innovation. It can be optimized from the following three aspects:

The first is to accelerate the cultivation of service providers for model implementation.Application innovation requires innovation entities to have a deeper understanding of vertical industries and scenarios. Especially when more complex scenarios and higher-level Agent forms are involved, industry understanding and scenario insight are particularly important to improving application effectiveness. At present, most model manufacturers have relatively limited industry knowledge accumulation and industry refinement service energy, and the AI Agent products launched are relatively insufficient in professional depth, making it difficult to meet the needs of professional and complex scenarios. Data shows that 70% of AI Agent products in GPTs stores have a life cycle of less than 30 days and are severely homogenized. In fact, it is believed that breakthrough innovation on the application side requires more diverse application service groups.

“Consider accelerating the cultivation of large-scale model implementation service ecosystems, encouraging more technology-based enterprises that understand both technology, industry insight, and business to participate in industry application innovation, promote traditional industry software service providers to accelerate business transformation, and actively participate in large-scale model implementation application services. rdquo; In fact, it is recommended that a number of high-quality implementation service companies and AI Agent implementation cases can be regularly selected and recommended to the society by industry, so that while fully activating the innovation vitality of large model applications, technological innovation can also benefit thousands of industries as soon as possible.

The second is to accelerate the opening of multiple scenarios and data.The realization of AI Agents ‘capabilities in industry applications needs to be driven by multiple scenarios and driven by professional data. It is suggested that we can give full play to the advantages of my country’s rich application scenarios and massive data resources, and accelerate the opening of scenarios and the release of the value of data elements.

In fact, it is proposed that at the scenario level, we can consider taking the lead in encouraging government agencies, key state-owned enterprises, industry leaders and other entities to open up more first-time scenarios in high-end manufacturing, medical, financial and other industries to stimulate the vitality of scenario-based application innovation. At the same time, we can create a technical testing and verification platform by industry to provide a public welfare platform for testing and evaluating the effectiveness of models by simulating real business scenarios, reducing technical verification costs; At the data level, we can consider having regulatory authorities take the lead in collecting professional field data of leading enterprises in various industries, using data sandbox tools to open up enterprises in an orderly manner under the principle that data is available and invisible, and optimize the application effectiveness of models.

The third is to accelerate the promotion of standard supply in key industries and key links.Artificial intelligence is a complex information system. For example, high-level AIAgent is a systematic integration of multiple models, multiple tools, and multiple devices. To achieve efficient operation and interaction of the system, it is necessary to unify various technologies and interfaces. Standards and requirements. Standards are crucial to enhancing the applicability of technologies and products, reducing information connectivity costs, and accelerating the application process of artificial intelligence technology. In order to accelerate the formulation of artificial intelligence industry standards, the Artificial Intelligence Standardization Technical Committee of the Ministry of Industry and Information Technology was officially established at the end of 2024.

“It is suggested that we should consider this as an opportunity to follow the principle of urgency first, highlight application orientation, and face key issues such as data quality and interconnection. We should take the lead in launching relevant artificial intelligence industry standards and specifications in cutting-edge directions such as multimodal and Agent, and build a series of evaluation tools around standards., effectively guide industrial development with standards, and help our country seize the right to speak in technology and seize strategic initiative in the international competition of cutting-edge technologies. rdquo; In fact, further stated. (Blue Whale News Hu Jiehujie@lanjinger.com)