SOCAN is moving to tackle a core problem posed by generative artificial intelligence to music creators: the absence of transparency about when an AI system uses or is influenced by a human-made work. The Canadian performing rights organisation said it has entered an exploratory relationship with Toronto-based Musical AI to test attribution tools intended to identify when creators’ music is used in AI-generated outputs.
What SOCAN is testing and why it matters
Generative AI systems can produce music that echoes or borrows from existing works, but the companies behind those systems generally don’t disclose what source material they used. SOCAN says that leaves creators unable to know when their work has been repurposed, which in turn hampers rights enforcement and payment.
Through the relationship with Musical AI — a Canadian company the organisation describes as having expertise in big data, AI, rights management and music industry experience — SOCAN plans to explore whether attribution technology can:
- show creators when their work has been used by AI systems, or has influenced fully AI-generated outputs;
- provide sufficient accuracy and reliability to support claims and compensation for such uses;
- deliver transparency without signalling permission for AI platforms to exploit works.
“This relationship is not permission for AI to use your music or a shift away from human creativity.”
SOCAN frames the project as a protective and proactive step to preserve members’ rights in an environment where creators currently have "no visibility at all" into how their music might be used by AI.
Scope and limits of the announcement
The announcement is an exploratory engagement, not a deployment. SOCAN emphasises the relationship is intended to assess whether attribution technology can be accurate and reliable enough to facilitate compensation mechanisms when AI systems use existing music. The organisation also stresses that the work is aligned with defending human creators’ roles and revenue streams.
| Issue | What SOCAN aims to achieve |
|---|---|
| Lack of transparency | Test tools that can identify when music is used or influences AI outputs |
| Inability to enforce rights | Explore whether attribution can provide evidence for compensation |
| Protecting creator revenue | Assess whether tools support future payment and valuation models |
SOCAN’s statement highlights broader concerns within Canada’s cultural sector — and internationally — about the opacity of AI training and generation. By testing attribution technology, the organisation is attempting to build technical evidence that could shore up legal and commercial claims by songwriters and publishers.
Implications for creators, platforms and policy
If attribution tools prove accurate, they could change how creators interact with AI platforms and how the industry adjudicates claims of unauthorised use. Practical, verifiable attribution could enable creators to demand payment, issue takedown notices or negotiate licensing arrangements when AI systems draw from their work.
But the announcement leaves open several practical questions that SOCAN will need to resolve as testing progresses: whether attribution can work at scale across diverse AI models, how to verify matches in contested cases, and whether attribution evidence will be sufficient for contractual or regulatory enforcement. SOCAN’s collaboration with a domestic AI firm positions the organisation to test solutions tailored to the Canadian rights ecosystem.
The project also signals to policy-makers and platforms that rights-holders are pursuing technological remedies alongside legal and regulatory ones. As governments consider frameworks for AI transparency and accountability, functioning attribution tools could become part of the evidence base informing regulation or industry standards.
For music creators worried about losing control, credit and compensation — concerns SOCAN describes as "valid" — the organisation’s effort to test attribution technology is a step toward restoring some visibility into how AI systems use human-made music. Whether the technology delivers a reliable, enforceable solution remains to be seen.
