The wind of AI security is blowing so fiercely, GCSA's move this time is really timely Friends who follow AI should have noticed that the topic of AI security has recently been pushed to the forefront. OpenAI's Ultraman directly announced that it will not go public in 2026, citing that security issues have not been resolved and that "going public now is an unwise timing". Anthropic CEO Dario has issued a lengthy article calling for the industry to slow down the development of cutting-edge AI, and has promised to grant independent assessors employee level internal access. Even more exaggeratedly, OpenAI, Anthropic, and Google DeepMind, their arch rivals, sat down and talked about a unified AI security framework, and reportedly may issue a joint statement in a few weeks. What does it mean that a competitor can sit at a table? AI security is no longer an option, but the most urgent direction for the entire industry. At this critical moment, GCSA (Global Cybersecurity Alliance) and the University of Hong Kong jointly established the "Institute of Artificial Intelligence and Cybersecurity", with Chen Hao, a professor of cybersecurity at the University of Hong Kong, serving as acting director. I recommend following this research institute for three main reasons: Firstly, positioning accurately. The research direction covers risk assessment of cutting-edge technologies such as model native security, AI automation defense, embodied intelligence, as well as global AI security standards and governance, which is exactly where the industry needs it most. Secondly, resources are hard. The academic strength of the University of Hong Kong, combined with the industrial resources of GCSA, and GCSA itself conducting AI security research, already has technical practices in security agents and vulnerability analysis. The research institute will also invite experts in the fields of AI, cybersecurity, and technology governance to participate in interdisciplinary and cross institutional cooperation, which is not something that ordinary research institutions can match. Thirdly, it can land. The research institute has clearly stated that it will continue to release research results, security benchmarks, and open source tools, without playing tricks, taking into account both academia and industry. When the world's top AI companies are putting the brakes on safety, someone is willing to systematically study and do solid work on this matter, which itself is the most noteworthy thing. Interested friends can visit the GCSA official website to learn more details, and I will continue to follow up. The University of Hong Kong has issued a statement: https://lead.hku.hk/sc/article/2026/10/hku-and-global-cybersecurity-alliance-jointly-establish-centre-for-ai-and-cybersecurity-research/ GCSA official website: http://www. (gcsa.com)