On September 22, 2026, a University of Alberta study landed a blunt finding: a Canadian musicians AI survey shows working artists feel squeezed by AI, streaming, and shuttered venues — and most see big tech as a direct threat to their livelihoods.
What the Canadian musicians AI survey found
The study, conducted by the University of Alberta and summarized by researcher Brian Fauteux, surveyed 263 working musicians across Canada and asked how they are navigating the current industry. According to the university’s report published on September 22, 2026, 87 percent of the 142 respondents who addressed generative AI viewed the technology negatively. Many called AI song generators like Suno antithetical to their definition of art. Others flagged environmental worries and said the industry is not taking those concerns seriously.
The survey describes a pincer movement. In the physical world, venue closures have thinned opportunities to build a fan base or make up income on the road. Online, discovery and distribution are concentrated in the hands of a few platforms, which sets the terms for reach and pay. Respondents described being both “personally and professionally” offended by generative tools that can replicate styles without consent, while also seeing fewer fair-paying channels to monetize original work.
Read at face value, the numbers capture anger. Read in context, they map a structural problem: if live and digital both stall, the middle of the market — the working-class musician — has nowhere to go.
How streaming and platform power compound the squeeze
The money math of streaming was already tight. Discovery favors catalog scale and constant releases, while per-stream rates hover low and vary by service. For Canadian artists, those conditions meet a shrinking local venue circuit. That mix turns experimentation and growth into luxury goods.
The Canadian musicians AI survey points to consolidation as a root cause. Artists no longer negotiate with a patchwork of regional gatekeepers; they accept algorithmic terms from a handful of global firms. When those firms push AI-generated tracks into recommendation buckets, even modestly, it can displace listening that once flowed to human creators. The fear is simple: more supply, same demand, worse odds.
Policy is inching into the space. Ottawa’s Online Streaming Act aims to modernize broadcasting rules for the platform era. Meanwhile, Parliament’s proposed Artificial Intelligence and Data Act (AIDA) sets a framework for AI accountability. But the survey’s portrait implies something tighter: rules that don’t move money or attribution won’t change daily life for working artists.
One practical lever is transparency. Clear labeling of AI-generated music in feeds and charts would let listeners choose and help rights holders contest misuse. Another is contributions tied to local outcomes: if large platforms benefit from Canadian audiences, direct funding for domestic artists and venues can refill the pipeline the closures drained.
AI music generators in the studio, and on stage
Generative tools are fast, cheap, and getting better. Their defenders see a new instrument. Many artists in this study saw a copier. The rift is not abstract. When style transfer and voice cloning ride on datasets scraped without permission, they collide with identity and craft.
The live side is blurring, too. The Guardian’s technology coverage on September 20, 2026 noted a UK startup raising $20 million to recreate gigs with “hyper-realistic” digital avatars (The Guardian). That idea will find an audience, but it also underlines a core anxiety in Canada: when virtual shows and text-to-music apps scale faster than fair pay systems, the bargaining position of human performers weakens.
For some creators, AI can be a collaborator or a sketchpad. For most in this sample, it reads as displacement. The Canadian musicians AI survey suggests a backlash rooted less in fear of tools and more in the absence of consent, credit, and compensation.
What could shift in Canada next
There is no single fix, but there are clear tests for whether policy and platforms are moving beyond rhetoric.
- Consent by default: Training on musical works and voices should require permission. AIDA can draw that line for high-impact AI systems, and rights groups can translate it into industry standards.
- Labeling that matters: Platforms should tag AI-generated tracks throughout the product, including recommendations and playlists. Labels need to be visible, not buried in metadata.
- Money where discovery happens: If feeds feature AI music, a portion of revenue should flow to funds that support human creators and the local venue network.
- Data for artists: Give creators access to granular recommendation and audience data so they can see when and how AI content competes with their tracks.
Artists can also push practical hedges. Lock in mailing lists and direct-to-fan channels that sit outside third-party algorithms. Diversify into sync and commissions where contracts specify use and credit. Team with peers to negotiate better local terms with venues that remain open.
The point isn’t to reject technology. It’s to rebuild the bargaining floor. As SOCAN’s own education materials on streaming payouts make clear, rate structures are complex and favor scale; without new guardrails, that complexity will tilt even further toward the largest catalogs.
The University of Alberta’s findings capture a moment of pressure and skepticism. They also describe the stakes if nothing changes. If venue closures continue, if platform consolidation deepens, and if AI music quietly floods discovery, the next Canadian musicians AI survey may show fewer working artists left to answer it. For more on this, see bloomberg.com and nytimes.com.
Related reading: NVIDIA • Meta AI • AI & Big Tech