The Center for Data Innovation recently organized a workshop in Beijing, hosted by the Institute for Artificial Intelligence at Peking University, to discuss the risks associated with open-source AI. The event took place on December 10-11, 2024, and included a panel of international experts.

During the workshop, participants collaborated to create a statement aimed at enhancing international cooperation to improve safety and security in open-source AI. Daniel Castro, director of the Center for Data Innovation, highlighted that "while many countries are competing to be global leaders in AI, they do not need to compete on developing the foundations for AI safety." He noted that existing open-source AI models present an opportunity for collaboration among nations. However, he pointed out that various barriers hinder some experts from contributing more significantly to global efforts in AI safety.

The statement developed during the workshop stresses broadening participation in global forums on AI safety and eliminating obstacles to international collaboration. It advocates for inclusive engagement across different countries and emphasizes the need for reliable testing methods and transparency in incident reporting. By focusing on high-impact risks, it aims to ensure responsible development and deployment of open-source AI technologies.

For further information or inquiries about this initiative, contact Nicole Hinojosa at [email protected].