As artificial intelligence continues to advance, one question lies at the heart of its evolution: where do we want it to take us?
At Mila, the frontier of intelligence is not a fixed boundary. It is where acquired knowledge meets what remains to be discovered—the space where a scientific breakthrough can unlock new possibilities for science, technology, the economy, society and humanity.
But pushing this frontier is not simply about developing more powerful systems. It also means understanding how they work, assessing their impact and ensuring they serve individuals and communities. It is at the intersection of scientific ambition and responsibility that Mila is shaping the next frontier of artificial intelligence.
Exploring the Unknown to Advance Knowledge
Every advance in AI begins with a question that science has yet to answer: How do models learn? How can we better assess their capabilities and limitations? How can we design more reliable systems that support discovery in fields as varied as healthcare, climate science and fundamental research?
This is the essence of the scientific frontier: deepening our understanding of artificial intelligence while using it as a new tool for exploration. By bringing together a major academic research community, Mila creates an environment where ideas move across disciplines and fundamental research can lead to breakthroughs extending far beyond computer science.
This culture of open research plays an essential role. Sharing knowledge enables more researchers, organizations and communities to contribute to scientific progress, validate results and build on existing discoveries. In doing so, it helps make AI a collective engine for discovery rather than a capability reserved for a select few.
Developing More Powerful, Open and Sustainable Technologies
The next generation of AI systems will need to do more than push the limits of performance. They will also need to become more accessible, efficient and responsive to real-world needs.
This is the challenge at the technological frontier. As models grow in size and complexity, their computing and energy requirements increase. Research must therefore advance on several fronts at once: improving system capabilities, reducing resource consumption and supporting open approaches that enable more people and organizations to understand, adapt and develop these technologies.
This research also strengthens technological sovereignty. A society with the talent, infrastructure and expertise required to design and evaluate its own systems retains greater control over the technologies shaping its economy and institutions.
Making AI a Driver of Economic Resilience
A discovery creates meaningful impact only when it moves beyond the laboratory, addresses a real need and is implemented responsibly. The economic frontier lies precisely at the intersection of research, innovation and adoption.
For organizations, integrating AI involves much more than simply adding a new tool. AI adoption begins with understanding where it can create value, then preparing data and teams so that its use supports a coherent strategy. When these conditions are in place, AI can improve productivity, accelerate innovation and strengthen the resilience of businesses and value chains.
Mila brings these worlds together by building bridges between research, industry and entrepreneurship. These connections give organizations access to leading scientific expertise, allow researchers to test their work against real-world challenges and help startups transform science-based ideas into solutions with the potential to scale.
The objective is therefore not to adopt AI at any cost, but to build an economy in which responsible intelligence can generate lasting benefits at every stage of the value chain.
Strengthening Trust at the Heart of Transformation
Artificial intelligence is already influencing how we work, access information, receive services and make decisions. Its development therefore raises questions that extend far beyond technology.
How can institutions use these systems while remaining transparent and accountable? How can we protect rights, languages and cultures? How can we prevent these tools from reproducing inequalities, excluding certain communities or diminishing individual autonomy?
The societal and human frontiers are inseparable from these questions. The first concerns the ability of institutions and public services to adapt, govern intelligent systems and preserve public trust. The second reminds us that dignity, rights and well-being must remain at the centre of technological decisions.
This vision turns trust and safety into genuine technical capabilities. They cannot be added at the end of a project; they must guide the design, evaluation and deployment of systems from the outset. Trustworthy AI therefore depends as much on scientific rigour as it does on dialogue among researchers, organizations, policymakers and affected communities.
Five Frontiers, One Shared Ambition
Scientific, technological, economic, societal and human: these five frontiers do not advance in isolation. A more powerful technology has value only if it can be understood and integrated. Large-scale adoption cannot be sustainable without trust. A more innovative economy cannot be truly resilient if its technologies undermine rights, cultures or the environment.
Shaping the next era of intelligence therefore requires bringing these dimensions together. By connecting fundamental research, open innovation, entrepreneurship, governance and society’s needs, advances in AI can expand the realm of possibility for the benefit of all.
Mila at ALL IN 2026
On September 16 and 17, Mila will co-present ALL IN 2026 and play a central role throughout the event. Its presence will offer participants several opportunities to discover the research shaping AI’s next frontiers, connect with the Mila community and explore new possibilities for collaboration.
On the ALL IN stages, Valérie Pisano, President and CEO of Mila; Hugo Larochelle (Université de Montréal and McGill University), Scientific Director of Mila; and Yoshua Bengio (Université de Montréal), Founder and Scientific Advisor of Mila, will deliver talks. Mila faculty members Laurent Charlin (HEC Montréal), Glen Berseth and Emma Frejinger (Université de Montréal) will also participate in panels exploring the scientific, technological and economic transformations driven by AI.
At the Mila Pavilion, the organization will host three areas for discussion and discovery:
- Sovereign AI and open technologies, exploring the capabilities societies need to maintain collective control over technological development;
- Trust and safety as technical capabilities, demonstrating how these principles can be embedded directly into system design;
- Scientist-entrepreneurs with Mila Ventures, exploring the journey from cutting-edge research to the creation of high-potential companies and solutions.
Entrepreneurship within Mila’s community will also be well represented, with seven Mila-affiliated startups featured among ALL IN 2026’s Top 100 Canadian AI Startups.
We invite all participants to visit the Mila Pavilion, attend talks and panels featuring members of its community, and meet the people working today to define where intelligence can take us tomorrow.
