Original title: 360 Group participated in the annual meeting of the Artificial Intelligence Security Governance Special Committee of the Cyber Security Association to share the security solution of the large model.
On December 14, the annual academic meeting (2023) of the Artificial Intelligence Security Governance Professional Committee of the Cyberspace Security Association of China was held in BeijingSH, hereinafter referred to as "360") Group was invited to share the security risks and governance solutions of the large model.
Ye Jian, Chief Operating Officer of 360, pointed out in his sharing that as an important technology to promote industrial digitalization, artificial intelligence large models have brought about development and are also facing unprecedented security challenges, and the security of large models has become the focus of attention of all parties. Ye Jian believes that the main security problems faced by the development of large models include: technical security issues focusing on network security and data security, content security problems caused by the "fabrication" ability of large models, prompt injection attacks, etc., and long-term security issues such as controllability.
As the only cross-border technology enterprise in China with both digital security and artificial intelligence capabilities, 360 has set out to overcome the security problems of large models, and has launched enterprise-level AI large model solutions in accordance with the four principles of "security, goodness, trustworthiness, and controllability" to support the training, deployment and delivery of vertical large models in nearly ten core scenarios.
Zhou Hongyi, the founder of 360 Group, once said that to solve the security problem of large models, we must make a breakthrough in technology, and we cannot rely on large model enterprises to consciously self-discipline, but also need security companies to devote themselves to research and find solutions.
As a leading enterprise of artificial intelligence large model, 360 actively promotes the ecological construction and technology development of domestic large models, and has been awarded the "Deputy Leader Unit of the Security Governance Committee" and the "Core Participating Unit" of the series of evaluation specifications for the safety risk prevention capability of large models by the Chinese Artificial Intelligence Industry Development Alliance, the leader unit of the "National Artificial Intelligence Standardization Overall Group Large Model Special Group" by the China Electronics Standardization Institute of the Ministry of Industry and Information Technology, and the "Large Model Working Group of the Key Laboratory of Artificial Intelligence" by the Academy of Information and Communications Technology of the Ministry of Industry and Information Technology At the same time, as the first batch of member units of the Artificial Intelligence Security Governance Professional Committee of the Cyberspace Security Association of China, he participates in the corpus and security evaluation of large models, and promotes the construction of a national standard system for large models.