National Summit on AI and Culture

In mid-March 2026, Artengine was invited to attend the National Summit on Artificial Intelligence and Culture at the Banff Centre for Arts and Creativity — Canada’s first federal summit dedicated to the intersection of AI and cultural practice. Convened by Ministers Marc Miller and Evan Solomon, with both ministers present for the full two days, the summit brought together approximately 300 participants: unions, publishers, broadcasters, streaming platforms, arts service organizations, universities, Indigenous cultural leaders, technology companies, and government representatives from multiple provinces and territories.

It is a rapidly shifting cultural landscape we are all working in and everyone feels like they are on unstable ground. Sometimes this is exhilarating. Sometimes it is terrifying.

Why We Were There
Artengine has been working at the intersection of art and technology for over 30 years. The framing that often emerged in discussion during the Summit was that AI, and technology generally, was something happening to culture. We are not passive recipients of technological change. We are an organization built on the conviction that artists and cultural workers are active participants in shaping how technology develops and what it means. Part of why we accepted the invitation and made the trip was to make this perspective visible in what can feel like a very polarized space.

The Shape of the Conversation
The summit was organized around three pillars: Build (adoption and opportunity), Protect (rights, intellectual property, labour), and Empower (skills and tools). Much of the most active advocacy energy came from the Protect side — unions, music rights organizations, publishers, and broadcasters making the case for stronger copyright protections against unauthorized AI training on creative work. These are legitimate concerns, and the scale of mobilization behind them is real.

From the outset, host Catalina Briceño articulated the many tensions in the room — from the cheering ‘pom-poms’ to the angry ‘pitchforks’ — and encouraged participants to find energy and generosity in the space between. We felt this was one of the summit’s quiet successes: energy drawn from that entangled space in between.

Two Pillars, Two Kinds of Stakes
One of the distinctions that crystallized for us at the summit is the difference between AI’s role in creative production and its role in operations — and the very different questions each raises.

In the production context — where AI tools are woven into the actual making of cultural work — the ethical stakes are high and clarity is not easy to come by. Questions about authorship, labour displacement, training data, and the nature of creative agency are sticky and messy. But it is important to remember that we have affordances and possibilities for experimentation that not all sectors of society do; health care or law are far more complicated with much higher stakes. In culture we have a chance to play an experimental role for society in exploring both dangers and opportunities in this messy space.

When a performing arts company uses real-time AI image generation as part of a live performance, we are watching a pipeline compression: work that might once have required multiple media artists over many weeks is produced in hours. What that means for creative practice, for the value of craft, for the artists whose aesthetic sensibilities were absorbed into training datasets — these are not questions to sidestep. But many new things can also be found in the process. The parallel of what video brought to art and culture is worth considering here. We want artists to be where new technologies are so that we can help shape their future.

In the operational context — AI as a tool for grant writing, research synthesis, communications, documentation — the ethical terrain is different, though not absent. Research cited at the summit found that AI tools compress performance gaps most significantly for mid-range performers. For a small organization like Artengine, or the dozens of two-to-five person nonprofits that make up the artist-run sector across Canada, this is not a small finding. The question is not whether we can compete with well-resourced institutions. It’s whether we can unlock the capacity that has always been there but constrained by the realities of underfunded cultural infrastructure.

Rather than thinking about small organizations finding efficiencies, it is more useful to think about unlocking capacity — the ability to do the thinking, the connecting, the creating, that administrative pressures of the sector have always crowded out.

One dimension of these questions that stayed with us: bringing twenty or more years of embodied practice to these tools is a fundamentally different experience than entering a field with AI already present. The summit was notably quiet on younger voices. What cognitive skills become more essential when certain kinds of processing can be offloaded? What kinds of knowledge — hard-won, contextual, embodied — become more valuable precisely because they cannot be replicated? These are questions we are sitting with, and we expect they will shape how we think about mentorship, programming, and organizational learning in the years ahead.

From Data Trust to Managed Emergence
We came to the summit with an interest in the idea of a cultural data trust — a governance structure through which publicly funded cultural organizations might collectively steward and benefit from the data generated by their work, rather than having it extracted by platforms or sitting inert in opaque institutional silos.

We still believe this is an important tool for cultural sovereignty, but the summit helped rethink how we might arrive there.

The conversations at Banff — especially those with Indigenous data sovereignty thinkers, researchers working on federated governance models, and practitioners from organizations like Hypha Worker Co-operative — reinforced something important: a top-down design creating a single national architecture for cultural data is unlikely to manifest quickly enough, and would be very hard to get right at scale. The diversity of the cultural sector — in scale, mandate, governance structure, regional context, and relationship to technology — is one of our strengths. The risk is that any top-down solution turns that diversity into inequity.

What we are starting to see, both in organizational imaginations and in practice, is something like a vision of managed emergence: federal and provincial governments, alongside the larger arms-length bodies, playing a role not as architects of a singular system but as coordinators and connectors — facilitating knowledge exchange, surfacing successful protocols and workflows from across the country, and creating conditions for iterative, distributed experimentation that can eventually cohere into something more durable.

There is also a real and underexplored opportunity in increased transparency from public funders about the data they steward. Canada Council for the Arts, for instance, has been gathering proposals, project reports, and sector knowledge for decades. That body of material — if made more navigable and more connected — could be a genuine resource for the sector rather than remaining in institutional storage. We don’t know what the infrastructure looks like internally, and this is not a criticism. It’s an invitation: what would it mean to treat that data as a shared cultural resource?

What We Brought Home
The summit closed with a formal announcement and a sharpened sense of what Artengine’s role in this conversation should be: not to speak for a sector, but to model a way of working — experimentally, critically, collaboratively — that others can learn from and connect to. We have been doing this for three decades. The urgency of the current moment doesn’t change that orientation. It reinforces it.