China launches AI system to accelerate scientific research workflows
The Chinese Academy of Sciences released ScienceOne 100, a suite of artificial intelligence models designed to speed up research across multiple scientific disciplines. The system launched in Beijing on Tuesday and includes eight discipline-specific models plus a foundational scientific model first deployed in July 2025.
The system addresses a practical problem facing large research facilities. The Beijing Spectrometer, a particle detector that has operated for 16 years, generates vast amounts of data that traditional analysis methods struggle to process efficiently. Liu Beijiang, a researcher at the Institute of High Energy Physics, said automated workflows guided by AI can reduce repetitive tasks and shorten analysis cycles.
How the system works
ScienceOne 100 uses large language models-AI systems trained on vast amounts of text and data-adapted specifically for scientific research tasks. The models can analyze academic papers, process experimental data, and generate material designs based on scientific literature and physical laws.
The system's most widely used feature is an intelligent literature assistant that analyzes academic papers. According to the Academy, it achieves 90 percent accuracy compared to about 70 percent in mainstream international systems, reducing errors where AI systems generate incorrect information.
Researchers can intervene during operations and adjust tasks in real time. The Academy reports this capability reduces research time by more than 60 percent.
Discipline-specific applications
The eight specialized models focus on different fields. A model for near-space research-the region between 20 and 100 kilometers altitude-simulates atmospheric and electromagnetic conditions to support communications, navigation, and remote sensing applications. Yang Yanchu, a researcher at the Aerospace Information Research Institute, said the model performs better than comparable systems in logical consistency and technical reasoning.
In materials science, the system integrates scientific literature, physical laws, and material properties to generate customized designs. Liu Jianjun, a researcher at the Shanghai Institute of Ceramics, said this approach speeds development of high-performance materials for aerospace and integrated circuits-work that traditionally takes about 15 years due to the number of variables involved.
China's broader AI strategy
The system aligns with China's 15th Five-Year Plan (2026-30), which emphasizes "AI Plus"-using artificial intelligence to process complex scientific data and improve research efficiency. Zeng Dajun, a researcher at the Institute of Automation, said the foundational model promotes open sharing of data, models, and algorithms across the scientific community.
Learn more about AI for Science & Research or explore AI Research Courses to understand how these tools are reshaping scientific workflows.
Your membership also unlocks:
