On September 12, Li Ke, Co-founder and CEO of Haitianyuya, was invited to participate in the "Data meets AI: Dual Engines of the Intelligent Era" forum at the 2025 Inclusion·Bund Conference. He delivered a keynote speech titled "Data Exploration and Practice in the Era of Large Models," focusing on new data paradigms in the age of large models and Haitianyuya's practical exploration and applications in the AI data field.
On September 11, the 2025 Inclusion·Bund Conference officially opened at Shanghai Huangpu World Expo Park under the theme "Reshaping Innovation and Growth." As one of Asia's most influential fintech summits, the Bund Conference attracts global attention with its openness, diversity, and forward-looking perspective.
The "Data meets AI: Dual Engines of the Intelligent Era" insight forum was jointly hosted by the Chinese Association for Artificial Intelligence, Shanghai Jiao Tong University, and Ant Group. The discussion centered on the question: "With human data available for large model training becoming increasingly scarce and Scaling Law gradually losing effectiveness, how can we break through the intelligence ceiling?" Multiple authoritative experts from industry and academia provided new solutions: data has driven AI development, while AI has also ushered in a new round of data evolution. The fusion of dual engines is the direction of progress.
During the conference, Li Ke delivered his keynote speech "Data Exploration and Practice in the Era of Large Models," sharing global AI data industry development trends from an industrial practice perspective and providing cutting-edge insights into data industry development in the large model era.
**High-Quality Dataset Construction Becomes New Breakthrough for Large Model Development**
Data, as the first engine of the intelligent era, is transforming from a supporting role to a core driving force. Li Ke pointed out that systematic construction and industrial application of high-quality datasets represent a new breakthrough for advancing large model development. He emphasized: "Future large models will pursue not just data volume, but a leap in data quality."
The data industry is undergoing a major transformation from labor-intensive to technology-intensive and knowledge-intensive. During his presentation, Li Ke provided detailed explanations of embodied intelligence data design and collection methods, demonstrating how to obtain data closer to the real physical world through motion capture and sensor fusion. He shared intelligent annotation processes for autonomous driving data, achieving efficient processing of complex traffic scenarios through platform-based tools, and introduced the production process of chain-of-thought datasets, emphasizing how annotating reasoning processes enhances the logic and interpretability of large models.
**Technological Innovation Drives Data Value Realization**
As the second engine, AI technology is profoundly changing how data is processed and utilized. Li Ke noted that the data industry is accelerating toward intelligence, engineering, and platformization. Haitianyuya's independently developed DOTS integrated data processing platform covers comprehensive engineering data services including data collection, cleaning, annotation, quality inspection, and management. It supports automated annotation for multimodal data, efficient data quality control, flexible project management, and data security and compliance assurance, ensuring clients can train higher-performance large models with high-quality multimodal data.
Looking ahead, only by achieving deep integration of data and AI, establishing comprehensive data standard systems and quality assessment frameworks, can we truly unleash the enormous potential of intelligent technology and drive the intelligent era to higher levels of development.
With its deep expertise in the data field and technological innovation, Haitianyuya is actively promoting standardized and intelligent development of the AI data industry. Moving forward, the company will continue to collaborate with industry partners, empowering various industries with high-quality data, jointly exploring new pathways for AI industry implementation, and contributing to building a more intelligent, efficient, and inclusive future society.