GBA lab for generative AI safety and development inaugurated

The Guangdong–Hong Kong–Macao Greater Bay Area Generative AI Safety and Development Collaborative Laboratory officially opens in Shenzhen’s Hetao Zone today. Photos by Lin Jianping

The Guangdong-Hong Kong-Macao Greater Bay Area Generative Artificial Intelligence Safety and Development Collaborative Laboratory was inaugurated in the Shenzhen Park of the Hetao Shenzhen-Hong Kong Science and Technology Innovation Cooperation Zone today.

The inauguration ceremony brought together representatives from government departments, leading AI companies, and specialized enterprises, including Tencent, OPPO, XPENG, and Sangfor Technologies, as well as higher education and research institutions such as Sun Yat-sen University, South China Normal University, Guangzhou University, and The Hong Kong University of Science and Technology.

A sign marks the entrance to the joint R&D and testing center inside the collaborative AI safety laboratory.

Guided by the National Computer Network Emergency Response Technical Team/Coordination Center of China (CNCERT/CC) and led by the Cyberspace Affairs Office of Guangdong Province and the Guangdong Center of CNCERT/CC, the collaborative laboratory represents an innovative model of governance cooperation. It was jointly established by government bodies, enterprises, universities, research institutes, and media organizations across the Greater Bay Area.

The lab aims to integrate resources from various institutions to build an innovative ecosystem that supports companies, industries, and safety governance. It strives to position the Greater Bay Area as a leading hub for the safe development of generative AI in China.

“The safety of large models is now a central concern within the AI industry,” said Zhou Peng, an expert from the laboratory and CTO of Luxi Tech, a Shenzhen-based company specializing in neuromorphic language models, during an interview at the ceremony.

A representative from Guangzhou Yundie Technology Co. briefs visitors on the company's self-developed data monitoring system designed for generative AI safety.

She explained that while the strength of generative AI lies in its ability to assist with questions from diverse fields, the answers it produces can carry risks, including ethical violations and factual inaccuracies.

“With the establishment of the collaborative lab, members can work together to develop unified standards, which will benefit both companies and the industry as a whole,” Zhou added.

In 2023, China introduced regulations for generative AI that require service providers to undergo mandatory model and algorithm safety assessments.

According to data, Guangdong Province is home to 92 registered generative AI services, ranking second in China.

A poster at the collaborative lab showcases Sangfor Technologies’ security-focused GPT model.

Qi Jiayin, an expert at the collaborative lab and a professor at the Cyberspace Institute of Advanced Technology at Guangzhou University, noted that many companies struggle to conduct these assessments. “The collaborative lab functions as a centralized, shared platform providing safety and compliance services for the entire industry. This public service model significantly reduces compliance costs for local AI companies, enabling them to redirect resources and capital from regulatory challenges toward their core mission — technological research and development,” she added.

Tian Xuesong, founder and CEO of Guangzhou Yundie Technology Co. Ltd., said that the lab will enhance the soft power of AI and large model enterprises in Guangdong, helping them build a distinct competitive edge both nationally and globally, and further boost the development of AI and large-scale model industries in the region.

Smart devices powered by Luxi Tech’s self-developed neuromorphic large model on display at the collaborative laboratory.