On the morning of August 17, 2026, the Shanghai High-Value Data and Corpus Development Symposium was held at the Shanghai Municipal Government. Zhang Ying, Deputy Secretary-General of the Shanghai Municipal People’s Government, chaired the meeting. Relevant officials from the Shanghai Municipal Commission of Economy and Informatization and representatives from six enterprises in related fields attended the symposium.

Grouphorse Group was invited to attend the symposium as an enterprise representative specializing in expert-driven, multilingual, cross-cultural high-value data and corpus. Zhang Li, President of Grouphorse Group, delivered a presentation on the Group’s business foundation, high-value data and corpus capabilities, representative project practices, and future development directions. Jiang Feng, Assistant to the President, also attended the meeting.

The symposium focused on the supply capacity, industrialization pathways, and application ecosystem of high-value data and corpus in the context of Artificial Intelligence (AI) development. Participating enterprises exchanged views based on their respective practices on topics including vertical-domain corpus development, expert resource organization, data product circulation, cross-border compliance, and industry collaboration.

From Language Services to High-Value AI Data and Corpus

During the presentation, Zhang Li introduced Group’s development path in recent years as it expanded into AI data and corpus services. Grouphorse has long provided services to the United Nations and other international organizations, as well as domestic and international institutional clients, and has built substantial resources and capabilities in multilingual, cross-cultural, and specialized Language Services. In March 2026, Grouphorse officially entered Yangpu AI Agent Industrial Park and further extended its existing expert resources, language capabilities, and complex project management system into the fields of high-value AI data and corpus and post-training data.

Grouphorse currently focuses on expert-driven, multilingual, and cross-cultural high-value data and corpus. Its business covers expert QA and instruction data in specialized fields such as healthcare, life sciences, law, and ESG, model preference and alignment data, and specialized data services including medical imaging, multilingual, and multimodal data.

These capabilities have been applied across a range of scenarios, including the Internet, Artificial Intelligence, gaming, and scientific research, and Grouphorse has accumulated practical experience in projects involving medical professional annotation, multilingual AI data, expert knowledge data, and multimodal data.

Zhang Li noted that Grouphorse does not focus primarily on general-purpose, homogeneous data processing. Instead, it places greater emphasis on corpus organization and quality control in scenarios with high professional barriers. The Group aims to transform its long-term accumulated language capabilities, expert resources, and international project experience into high-value data capabilities for large-scale AI model training and post-training.

Municipal Leaders Inquire About Grouphorse’s Differentiated Development Path

During Group’s presentation, Deputy Secretary-General Zhang Ying asked about the Group’s current workforce size, long-term business experience, and the relationship between its AI corpus business and traditional Language Services, and further discussed with Zhang Li the foundation on which Grouphorse has developed differentiated capabilities in the high-value data and corpus field.

Zhang Li explained that Grouphorse did not enter the market through general-purpose data annotation. Instead, it leverages multilingual capabilities, expert resources, and complex project management experience accumulated through years of serving international organizations, research institutions, and specialized clients, with a focus on developing expert-driven, multilingual, and cross-cultural data and corpus.

In particular, in specialized fields such as medicine, law, life sciences, and ESG, Grouphorse emphasizes the participation of domain experts in task definition, standard development, difficult-sample processing, and quality review. Through this “expert-first” approach, the Group enhances the professionalism, consistency, and traceability of its data and corpus.

The meeting also covered the Group’s next-stage development, including deeper cultivation of vertical domains, data product circulation, and cross-border business. The meeting proposed that high-value data and corpus enterprises should further focus on specialized fields, pursue refined development, and explore mechanisms for connecting high-quality data and corpus products with data trading platforms.

For enterprises with an established international business foundation and cross-border service capabilities, the meeting also proposed making further use of platform resources such as the Lin-gang Special Area to study cross-border data compliance and related business development.

Regarding high-quality data and corpus and related Skill products, the meeting also proposed exploring connections with platforms such as the Shanghai Data Exchange and studying feasible pathways for data productization and market-based circulation.

These discussions provided new ideas for Grouphorse’s subsequent efforts to deepen its capabilities in vertical domains, improve its high-value data and corpus product system, and expand its cross-border data-related business.

“Expert-First” Approach Builds a High-Value Data and Corpus Quality System

The core value of high-value data and corpus lies not only in data scale, but also in professionalism, accuracy, and consistency.

With this in mind, Grouphorse highlighted its “expert-first” production and organization model at the symposium.

At the project initiation stage, domain experts participate in task definition, standard development, and sample calibration. During production, quality is continuously calibrated through trial annotation, sampling inspection, review, and difficult-sample processing, while a full-process quality control and traceability mechanism is established.

At the same time, drawing on its long-term accumulated resources of university, industry, and professional talent, Grouphorse can dynamically form expert teams according to the needs of different projects, and organize and deliver work in tiers based on professional requirements, task complexity, and quality standards.

This model is designed to address several core issues in high-value data and corpus production: who defines the standards, who determines professional quality, how quality can be replicated consistently, and how expert knowledge can be transformed into scalable and deliverable data products.

Deepening Specialized Data and Corpus Capabilities and Contributing to Shanghai’s AI Data Ecosystem

Zhang Li stated that being invited to participate in the Shanghai High-Value Data and Corpus Development Symposium was, for Grouphorse, both an opportunity to gain a deeper understanding of industry policies and development directions and an important opportunity to exchange practical experience with government departments and peer enterprises.

At present, the Artificial Intelligence industry is moving from simply having “data available for use” toward having “high-quality data available for use.” As large-scale AI models enter stages of specialization, verticalization, and agent-based development, expert-driven, highly reliable, and traceable high-value data and corpus will become important infrastructure.

Looking ahead, Grouphorse will continue to leverage its accumulated strengths in multilingual capabilities, specialized fields, international resources, and expert organization. Focusing on expert-driven, high-barrier, and differentiated development, the Group will continue to enhance its capabilities in AI post-training data and corpus, specialized knowledge data, and multimodal data, while actively participating in the development of Shanghai’s AI data and corpus industry ecosystem and providing more professional and reliable data and corpus support for the high-quality development of the Artificial Intelligence industry.