Security of large models crucial to bolster AI application, says head of cybersecurity firm

Safeguarding the security of large models is of vital importance to bolster the application of artificial intelligence across various sectors, said Hu Zhenquan, president of Chinese cybersecurity company 360 Digital Security Group.

New and complicated cybersecurity challenges, such as data leakage and malicious misuse of data, have emerged along with the integration of AI with a wide range of fields, he said.

Hu made the remarks at the ongoing 2025 World Internet Conference Wuzhen Summit, calling for efforts to jointly solve the security issues brought about by AI.

The company released a white paper on the security of large models, which systematically elaborates on key risks in large model operation and proposes security solutions, on Saturday.

The solutions not only focus on strengthening external security protection for models, but also emphasize platforms' endogenous security, providing a safe environment for the application of AI in fields like government affairs, finance and manufacturing, Hu said.

AI security governance remains a long-term and challenging task that requires concerted efforts across the entire industry, highlighting that the company will work with industry partners to build a solid AI security defense system to promote the safe, steady development of the AI industry.