World Data Organization Launched, with iFLYTEK as the Only LLM Company Participating in Its Founding

World Data Organization Launched, with iFLYTEK as the Only LLM Company Participating in Its Founding

On March 30, the World Data Organization (WDO) was officially launched in Beijing. As the world’s first specialized international organization dedicated to advancing data development and governance practices, WDO is headquartered in Beijing and guided by the mission of “bridging the data divide, unlocking data value, and driving the digital economy.” It has already brought together more than 200 members from over 40 countries, spanning 14 sectors including industry, finance, healthcare, and education.

As one of the organization’s seven founding members, iFLYTEK has officially assumed its role and stands as the only LLM company among them. Liu Qingfeng, Chairman of iFLYTEK, attended the inauguration ceremony.

Why an LLM Company Matters in Global Data Governance

For decades, the world has had the WTO for trade and the IMF for finance. But when it comes to data—one of today’s most strategic factors of production—there has long been no dedicated international mechanism to coordinate global action. Different approaches to data ownership, privacy protection, and cross-border data flows have created widespread data silos. As Tan Tieniu, Chairman of the World Data Organization, observed, the defining feature of the current wave of artificial intelligence is that it is data-driven. How to fully unlock the value and potential of data is a shared challenge facing all humanity. The founding of WDO is aimed precisely at closing this gap.

The rapid rise of large models has made this challenge all the more pressing. A growing reality across the industry is that simply piling on more parameters and computing power is no longer enough to drive meaningful gains in model capability. The next stage of progress will increasingly hinge on the ability to secure and govern high-quality data.

This means that data governance cannot be shaped at the level of theory and policy alone. It also requires deep participation from frontline companies with real-world experience in managing data at scale. That is precisely why the presence of a large-model company among WDO’s founding members matters.

 From Standards Development to Real-World Validation: What Underpins iFLYTEK’s Strength?

iFLYTEK’s role as a founding member is backed by years of systematic capability building around data.

As early as July 2024, iFLYTEK played a leading role in the release of ISO/IEC 5259-4:2024, the international standard on AI data quality. Four years in the making, the standard covers the full data lifecycle, from data acquisition and preparation to labeling, evaluation, and use. In March 2026, iFLYTEK also took the lead in a national key science and technology project under China’s “New Generation Artificial Intelligence” initiative, titled “Content Safety Governance Based on Data Value.” The project focuses on two core issues: the causal mapping between data value and model safety, and full-lifecycle security evaluation with tiered and classified governance.

Beyond standards and research programs, the strongest proof lies in large-scale validation in real-world scenarios. iFLYTEK’s General Practice CDSS is already operating on a normalized basis across 806 counties and districts in China, having generated nearly 1.1 billion assisted diagnostic recommendations. Every time a primary care doctor accepts or corrects a recommendation, real clinical experience is fed back into the model, forming a “data flywheel” in which high-quality data and model capabilities continuously strengthen each other. This has propelled the iFLYTEK SPARK Medical Model to chief-physician-level performance in multiple specialties. In education, authentic teaching data from frontline master teachers is also driving the ongoing advancement of the iFLYTEK SPARK Education Model, leading to the world’s first highly humanlike virtual teacher capable of explaining why mistakes happen.

Full-Stack Self-Reliance Makes Data Security a Real Technical Capability

The greater the value of the data flywheel, the higher the bar for security. Especially today, as AI applications represented by personal AI agents continue to gain momentum, sensitive data such as medical histories, academic records, and financial information is being tapped in unprecedented depth. Data security has therefore become a core issue that global governance can no longer afford to sidestep.

As one of the very few companies in China to achieve Full-Stack Self-Reliance in large models, iFLYTEK has built the entire technology chain—from computing infrastructure and training frameworks to core algorithms and model deployment—without external dependencies. This means that when large models are applied to highly sensitive domains such as healthcare, education, and government services, the boundaries of data security can be clearly defined and effectively controlled, rather than left as a hollow promise.

The establishment of the World Data Organization marks a new stage in global data governance—one that moves from fragmented approaches toward coordinated collaboration. As a founding member, iFLYTEK will uphold the principles of consultation, collaboration, and shared benefit, actively contribute to the development, alignment, and mutual recognition of data-related rules and standards, help break down barriers to cross-border data flows, and join hands with partners around the world to build a new order for data cooperation.