Where Are the Humanities at the AI Table?
Written by Tamy Emma Pepin
November 19, 2026
Canada is at a critical juncture in its artificial intelligence ambitions. Announced in October, the federal government's new National Council on Artificial Intelligence has been tasked with advising on how to turn Canadian innovation into stronger companies, better public services, good jobs and greater economic security. It’s an important mandate, supporting the implementation of Canada’s AI Strategy, at a time when Canada's position in the global AI landscape is becoming increasingly uncertain.
In his article L’IA pour tous : Faisons du transfert une priorité nationale, Ludovic Soucisse, CEO of Québec’s CCTT network, highlights the importance of investments in transfer and adoption, drawing attention to Canada's position in Tortoise Media's Global AI Index. Where Canada competed with the world’s top 3 global AI leaders in the early 2020s, the country fell from fourth to eighth place between 2021 and 2024, tying for the largest decline among countries that remained in the top ten throughout that period.
The C.D. Howe Institute's 2026 report, From Hype to Output, by Rosalie Wyonch, also points to the industry challenge: strong AI research performance but weaker commercialization, and lagging business adoption. Canada does not lack scientific talent. It does not lack engineers, mathematical prowess, world-class researchers or technological ambition. Could we be failing at social innovation?
What History Teaches Us
Across civilizations and centuries, some of the greatest periods of scientific discovery, invention and wealth creation have emerged when the humanities were in dialogue with the sciences. Enlightenment periods offer compelling examples.
The Italian Renaissance brought together artists, philosophers, architects, mathematicians and scientists. The boundaries between disciplines were considerably more fluid than the institutional divisions we recognize today. Far from being simply a period of extraordinary artistic achievement, the Renaissance transformed how human beings studied nature, understood their place in the world, and connected intellectual inquiry with practical invention.
Centuries earlier, during the Abbasid Golden Age, the mathematician and astronomer al-Khwarizmi helped establish the foundations of algebra. The word algorithm comes from his name. Al-Khwarizmi worked in ninth-century Baghdad, when Baghdad was a center of science, culture, arts and invention. There, he was the Astronomer-in-Chief of the House of Wisdom, an important cultural place of scholarship, interdisciplinary translation and intellectual exchange. Greek philosophical and scientific works were translated into Arabic, while Indian mathematics, Persian scholarship and other intellectual traditions contributed to advances in astronomy, medicine, mathematics and philosophy. I spoke about Al-Khwarizmi in a keynote on community building at Montréal’s PHI center in 2019.
These periods of extraordinary scientific and economic advancement teach us that scientific discovery, adoption and knowledge transfer depend on more than engineering and technical ingenuity. They flourish when the humanities and sciences meaningfully engage with each other, bringing together diverse intellectual traditions, perspectives, skillsets and disciplines within institutions and systems that enable invention, creativity, development and application of knowledge.
From Model Development to Governance
These historical examples are relevant to Canada’s technological ecosystem. During the formative years of Montréal's AI ecosystem, institutions such as Mila advanced groundbreaking research, while companies such as Element AI sought to transform research into commercially viable products. However, throughout this important chapter of Canada's AI development, an opportunity was missed: embedding the humanities, social sciences and creative thinkers upstream alongside the mathematicians, computer scientists, researchers, executives and product developers building AI.
While subsequent initiatives engaged the cultural sector, they often came after engineering, focusing on the downstream consequences of artificial intelligence on the sector, rather than involving the arts and humanities in shaping technological innovation. The composition of Canada's newly formed National Council on AI risks repeating this silo at the governance level.
A Pivotal Moment for AI’s Next Chapter
At the recent Attention Conference in Montréal, co-hosted by McGill University and the Centre for Media, Technology and Democracy, Valérie Pisano, CEO of Mila, mentioned that: "AI is not happening to us. It is a business product, shaped by decisions and choices. We retain full agency over what gets built and how it is deployed.”
She also emphasized that AI should expand human creativity, connection and wisdom, not just drive productivity for its own sake. If AI is fundamentally a product shaped by human choices, then the capabilities required to define its purpose, design it, translate research into applications and deploy it successfully are not purely technical. These are precisely the capabilities cultivated through the humanities, interdisciplinary thinking and the creative practice of polymaths. Like other skills, they can be learned and cultivated through specific approaches, methodologies and intentional practice.
At the same conference, Yoshua Bengio discussed the work of LawZero to address AI misalignment and emerging risks. Alongside Camille François and Maria Ressa, he emphasized the importance of international coordination, digital sovereignty and democratic governance, so that AI's trajectory is not determined by commercial pressures alone. In the Financial Times, addressing the global wave of security issues and the proliferation of AI agent hacks, he challenged the industry's position that recent attacks could be addressed through cybersecurity measures alone, arguing that the underlying problem lies instead in how AI systems are trained:
“At the core of all of these incidents is misalignment. AI agents are choosing goals that are misaligned with those of humans, and it is a consequence of their training.”
He later added on LinkedIn:
“Until we solve the alignment problem at its source, AI models will continue to optimize for the wrong goals.”
Bengio’s technical warning underscores why technical alignment cannot be solved by computer science alone. Before an AI system can be mathematically constrained to align with human goals, those goals, values, and ethical boundaries must be rigorously defined. While technical talent can provide the mathematical architecture for alignment, the humanities, alongside independent thinkers and researchers working outside traditional institutions, can contribute valuable worldviews, forms of reasoning and approaches that can help articulate values and frameworks.
Both Bengio’s and Pisano’s interventions point to two consequential layers of artificial intelligence: First, the belief systems, assumptions and intellectual frameworks guiding the design and training of AI models. Second, the human agency, leadership approach, and policy structures through which AI is directed, governed and translated into economic, territorial and societal outcomes. The question, then, is not merely how artificial intelligence is developed and governed, but how we connect knowledge across disciplines, and whose perspectives are allowed to shape innovation and adoption.
On Framing Problems and Definitions
Canada’s National AI Strategy identifies global competition for technical talent as a primary barrier to building domestic AI capability. While technical research is undeniably vital, technology cannot be built on technical talent alone. Other disciplines and experiences bring essential capabilities to identifying opportunities, diagnosing root problems, and co-designing enduring solutions.
Consider the word technology. Derived from Greek technē (5th century BCE) – meaning art, craft, or skill – the term has become predominantly associated in modern discourse with software, hardware, and digital infrastructure. In parallel, our understanding of innovation has constricted, privileging technological and digital transformation. Yet some of humanity’s greatest innovations were not technological products, but social inventions. Democracy is a prime example.
Our understanding of diversity would similarly benefit from a strategic refresh. In public policy and institutional frameworks, diversity has frequently been associated with identity politics and demographic categorization, most notably race. Yet race, and more fundamentally, the belief in race(s) plural, has no scientific basis; it is a social construct.
Once embedded in public institutions and policy frameworks, categories can acquire enduring social consequences that reinforce the very beliefs and divisions they seek to overcome. This raises a fundamental question about the nature of system inputs, including the assumptions we encode into our institutions, and how they dictate economic and societal outcomes.
These questions extend directly to intellectual diversity. Which forms of knowledge, reasoning, and creative practice are recognized within the institutions shaping what gets built and how, and which remain excluded because they do not fit neatly into established categories, organizational functions or hierarchies?
In 2019, I inaugurated ORB, a self-funded interdisciplinary initiative connecting arts, humanities, sciences, and well-being on Montréal’s MIL Campus. The prototype successfully brought together creative practitioners, researchers, artists and technologists through a range of activities that promoted connectivity, dialogue and interaction, across otherwise siloed spaces.
Quickly self-sustaining before pandemic disruptions, ORB presented a new model while exposing a fundamental flaw in Canada’s funding ecosystem. At the time, technology programs deemed the initiative insufficiently technological, while arts programs were heavily skewed to support digital and interactive media. ORB existed between categories – and that was precisely the point. Knowledge transfer and technological adoption, by nature, happen in the in-betweens. The lesson extends well beyond any single experiment.
Social innovation underlies technological innovation
Social innovation is the living soil in which technological innovation takes root. It improves networks of exchange, changes how institutions collaborate, and creates the foundational conditions for discoveries to emerge, technologies to be adopted, and their benefits to be shared. Yet social innovation remains severely underleveraged in public policy.
As Canada enters the age of sovereign AI, integrating social innovation directly into national economic strategy, business and institutional practices, and technological development is essential to reduce vulnerabilities, strengthen economic resilience, accelerate useful and meaningful innovation, rebuild public trust and support adoption. An AI ecosystem anchored in democratic values cannot be developed in verticals.
Achieving this requires a broadening of Canada's definition of innovation, to integrate social innovation more boldly alongside technological capability, recognizing it as an innovation practice in its own right, and encouraging its application across sectors and organizational functions.
Crucially, social innovation must extend more vibrantly beyond the non-profit sector and directly into the commercial engine of the country. Small and medium-sized enterprises (SMEs) represent 99.8% of Canadian businesses, yet they face severe structural adoption bottlenecks and must be better included in the design of AI solutions across priority sectors, moving from passive end-users to co-creators. SMEs should be able to develop and apply social innovation to their company culture, products and business models in the same capacity as they have access to technological implementation support.
Aligning public policy, funding and evaluation frameworks across sectors would also properly recognize the vital contribution social innovation makes to organizational capacity, value creation, community building, knowledge transfer, regional cultural vitality, and sustainable economic progress.
This shift also offers a tangible opportunity to reimagine diversity and inclusion. By evolving traditional corporate DEI initiatives into dynamic Social Innovation Programs, institutions can dismantle vertical silos while cultivating true intellectual diversity, polymathic reasoning, and innovating with new methodologies and approaches that make us and our land collectively stronger, safer, healthier and more prosperous.
Research laboratories, scale-ups, and governing bodies can deliberately foster creativity, imagination, and humanistic wisdom alongside engineering expertise by investing in human capabilities and interdisciplinary participation, through different formats of learning, collaborative activities and executive leadership, putting social innovation into practice from early-stage R&D through to commercialization.
Sovereign AI also requires interdisciplinary representation at the highest levers of governance shaping Canada’s AI future alongside financial and technological leadership, rather than confining creative faculties to cultural or specialized advisory committees, ensuring our technology remains anchored in human flourishing.
With our extraordinary wealth of research, talent, and technical infrastructure, Canada is uniquely positioned to translate its intellectual capabilities into AI solutions that optimize resource management and create lasting value for our economy, living ecosystems, and future generations.
What if Canada’s AI strategy aimed not merely to improve organizational efficiency, but also helped to usher in a new period of Enlightenment – one defined by polymathy, art, enterprise, and scientific discovery?
Tamy Emma Pepin is an independent researcher and the founding manager of UPPL, a strategic communications and development firm established in Canada and operating internationally.