Dr Zeynep Engin, Chair and Director of Data for Policy CIC, joined a delegation of international experts in Beijing to engage in dialogues focused on fostering collaboration around Open-Source AI Safety.
The two-day workshop, held on 10–11 December 2024 under the Chatham House Rule, was hosted by the Institute for Artificial Intelligence at Peking University and facilitated by the Centre for Data Innovation (CDI). This format encouraged open and constructive discussions, highlighting opportunities for international cooperation in advancing safe and transparent AI technologies. Following the event, CDI published a statement summarising the broad principles of collaboration developed by the participants.
Dr Engin’s contributions centred on the need to clarify the definition of “Open-Source AI,” highlighting its connection to broader uncertainties surrounding the overarching goals and practices of the AI field. Ranging from the release of model weights alone to more comprehensive approaches to openness—encompassing entire algorithmic systems, data, computing and development environments, and human, environmental, and social contexts—these vastly divergent interpretations and practices create significant barriers to collaboration across disciplines, sectors, and geographies, while also clouding public understanding and expectations.
Reflecting on the discussions, she emphasised the critical importance of fostering dialogues like these to bridge the polarising discourses surrounding global AI technology development. Rather than allowing divergent viewpoints to hinder progress, she underscored the need to build a shared infrastructure that fosters international cooperation while empowering local and regional innovation, particularly in resource-constrained environments such as the Global South. Such collaboration would ensure that AI advancements are not driven solely by competitive or conflicting interests but are instead aligned with the broader goal of benefiting society as a whole. By coming together across disciplines, sectors, and geographies, stakeholders can collectively develop ethical, transparent, and inclusive AI solutions that address global challenges and serve the common good.
As a leading global expert community for data- and AI-driven innovations in the public interest, Data for Policy CIC is eager to expand its collaborations across Asia. This initiative builds on the success of its first regionally focused conference in Hong Kong in December 2022, which explored the theme “Ecosystems of Innovation and Virtual-Physical Interactions.” The Hong Kong conference was part of a broader series, with complementary editions held in Seattle and Brussels.