Sunrise Spark Large Model First to Obtain National Security Standard Certification, Marking a New Milestone in AI Security!

Recently, the Xunfei Spark Large Model Open Platform successfully passed the pilot verification of the national security standards and became one of the first units to obtain the "Artificial Intelligence Security National Standard Compliance Self-Assessment Certificate". This certification marks another important step for Xunfei in the field of safety for generative AI.

This verification was based on GB/T45654-2025 "Cybersecurity Technology Basic Requirements for Generative AI Service", which is China's first specialized cybersecurity technical standard for generative AI. The standard clearly defines the basic security requirements that generative AI services must follow, covering training data security, model security, and various security measures, aiming to improve industry security levels and prevent major security risks.

Xunfei not only participated in the development of this standard throughout the process but also contributed its efforts during the pilot verification. To meet the standard requirements, Xunfei established a comprehensive security protection platform based on its practical experience in the security defense of the Spark Large Model. This platform includes dozens of technologies and provides "out-of-the-box" protection in multilingual and multimodal scenarios. These security measures ensure the security of the Spark Large Model during multiple iterations and daily operations.

In addition to security protection, Xunfei has also developed a new generation of implicit watermark technology for generative text. This technology can provide a unique digital identity for AI-generated content. Through implicit watermarks, users can effectively identify and regulate AI-generated content, further enhancing the credibility of generative AI.

As AI technology continues to advance, security risks are becoming increasingly prominent. In response to this situation, Xunfei will continue to prioritize user rights and experience, explore more technical solutions, and ensure that the development of generative AI proceeds on a safe and reliable track. In the future, Xunfei hopes to safeguard the safe development of AI, making every AI-generated content more secure and trustworthy.