On March 26, Mozilla announced the establishment of a strategic research partnership with Mila (Mila – Quebec Artificial Intelligence Institute) in Quebec, Canada, aiming to jointly promote the development of open source artificial intelligence and “sovereign AI” capabilities. This is the first time that Mozala has reached such an in-depth cooperation with a large artificial intelligence research laboratory. It is regarded by both parties as a landmark strategic collaboration and will carry out long-term and scalable joint research around "trustworthy" and "controllable" artificial intelligence.
Mozilla pointed out in the announcement that future artificial intelligence should not only be in the hands of a few countries or large enterprises, but should "belong to all mankind." To realize this vision, AI systems must be open, trustworthy, and built in a way that gives individuals, institutions, and countries real choices. Based on this judgment, the first phase of cooperation projects between the two parties will focus on the intersection of "trust and usability", including research directions such as privacy memory architecture for AI agents.
Mozilla believes that just as open source software stacks have greatly promoted the democratization of innovation and accelerated iteration in the Internet era, similar opportunities exist in the field of artificial intelligence today. At present, a large number of basic components have been formed in the open source community, from computing power and models to data and development experience. However, there are still obvious gaps in the key aspects that determine whether AI is "trustworthy, private, and adaptable to multilingual, multicultural, and diverse social organization methods." Mozilla hopes that by completing these links, open source AI can become a truly optional and reliable solution for governments, institutions and developers.
"We are working hard to build an AI future based on openness, privacy and humanity." Mozilla President Mark Surman said that this cooperation is a "delivery tool" to realize the above vision and is expected to bring new breakthroughs to governments, developers and enterprises. He emphasized that Canada is fully qualified to play a leading role in "AI sovereignty", and the cooperation between Mozilla and Mila is an important practice in promoting the implementation of this goal.
Mila emphasized that Canada has comprehensive advantages in the field of "trustworthy cutting-edge AI". Mila President and CEO Valérie Pisano pointed out that Canada has a deep accumulation of research, a value orientation and a willingness to "do AI differently." In her view, the focus of the next stage of competition in artificial intelligence is not only capabilities, but also “credibility”; Canada has unique leadership potential in both aspects. She emphasized that open and trustworthy AI is not a compromise on ambition, but rather sets a higher standard.
At the specific cooperation level, Mozilla will leverage its long-term experience in the open source field, active developer community, and ecological ability to transform research results into widely disseminated products. Mila brings world-class scientific research depth and experience in turning basic breakthroughs into practical tools and technology diffusion mechanisms. Both parties hope to prove that open source AI is fully capable of bridging the gap between cutting-edge research and real-world applications.
Both parties stated that an important goal of the cooperation is to develop technologies and methods that reduce reliance on closed systems and leave greater space for transparency, accountability mechanisms and shared innovation. At the geopolitical level, this partnership is also seen as a testing ground for promoting collaboration between “middle countries” in the field of AI. Historically, open source projects have long provided a basic framework for cross-regional and cross-jurisdictional technical collaboration, and this tradition will continue in the AI cooperation between the two parties. Mozilla and Mila also welcome the participation of more research institutions, developers and organizations with similar values to jointly improve the open source AI technology stack.
Mozilla said that this is just the beginning of a series of collaborative projects planned by both parties, and that "continuous and expanding" joint research work will be formed in the future. More information about Mozilla’s overall open source AI strategy can be found in its “Open Source AI Strategy” statement. The public can also visit Mila’s official website to learn more about this Canadian AI research institution.