A team of researchers from the University of Science and Technology of China and the Zhongguancun Institute of Artificial Intelligence has developed SciGuard, an agent-based safeguard designed to control the misuse risks of AI in chemical science. By combining large language models with principles and guidelines, external knowledge databases, relevant laws and regulations, and scientific tools and models, SciGuard ensures that AI systems remain both powerful and safe, achieving state-of-the-art defense against malicious use without compromising scientific utility. This study not only highlights the dual-use potential of AI in high-stakes science, but also provides a scalable framework for keeping advanced technologies aligned with human values.

ITU’s AI Skills Coalition and the Giga Initiative work to bridge the digital skills gap in African schools

A study published in Journal of Railway Science and Technology developed a class of polymer fiber-reinforced concrete that mitigates brittle behavior under low vacuum conditions. Using acoustic emission techniques, the research examined how low vacuum environments, fiber type, fiber content, and coarse aggregates affect the mechanical properties of two fiber-reinforced concretes, identifying an optimal material combination.

The 12th Heidelberg Laureate Forum (HLF) has come to a close. This year’s HLF took place from September 14 to 19 and brought together 28 Laureates of the most prestigious prizes in mathematics and computer science as well as 200 of those disciplines’ brightest Young Researchers of the next generation.

Microscopy is a vital tool for exploring the mysteries of life, enabling scientists to observe the dynamic changes of cells and organelles across spatial and temporal scales. However, because cells are primarily composed of water and exhibit low refractive-index contrast, it remains challenging to clearly visualize their internal structures. In recent years, Intensity Diffraction Tomography (IDT), an emerging label-free 3D imaging technique, has attracted increasing attention. Instead of relying on fluorescent dyes, IDT directly exploits the intrinsic refractive-index distribution of cells for imaging, thereby avoiding phototoxicity and photobleaching and making it particularly suitable for long-term live-cell observation. When combined with 3D fluorescence microscopy, this dual-modal approach can simultaneously provide molecular specificity and global structural information, offering a powerful tool for deciphering cellular components and their interactions. Nevertheless, long-term imaging (from several hours to days) is often hindered by the optical complexity of cells, environmental thermal drift, and mechanical instabilities of the microscope, which lead to time-varying aberrations and focal drift. Achieving stable, high-quality tracking of subcellular structures over extended periods therefore remains a key challenge.