Fair Dealing Week 2026: User rights then and now

Copyright's AI Crossroads

Fair dealing, a cornerstone of Canadian copyright law, allows limited use of protected works without permission for purposes like research, education, and criticism. Similar to fair use in the US, it balances creator rights with public access to knowledge. In 2026, this doctrine faces unprecedented scrutiny amid the explosive growth of generative AI technologies.

Courts in 2025 began delineating boundaries, with rulings like those in Anthropic cases deeming AI training on books 'transformative' under fair use, absent infringing outputs. Yet, judges like those in music publisher suits against AI firms expressed concerns over threats to creative markets. Over 70 lawsuits now pend, shifting focus from training data to AI-generated outputs that mimic copyrighted styles or content.

Real-world impacts hit creators hardest. Authors, musicians, and artists allege unauthorized scraping erodes revenues; a $1.5 billion settlement in Bartz v. Anthropic underscores financial stakes. AI developers counter that broad fair use is essential for innovation, warning licensing mandates could slow progress in a $1 trillion industry. Educators and researchers, key beneficiaries of fair dealing, worry stricter rules might limit access to AI tools for analysis and teaching.

Concrete deadlines loom: South Korea plans Copyright Act changes by mid-2026 permitting AI training with opt-outs and payments. The EU advances frameworks requiring disclosures and fair remuneration, rejecting flat-rate licenses. In the US, bills like the TRAIN Act enable subpoenas for training data details, while NO FAKES Act targets deepfakes. Inaction risks creators losing billions in uncompensated value, or AI firms relocating to lax jurisdictions.

Non-obvious tensions include copyright's dual role: empowering individuals against tech giants, yet potentially entrenching monopolies by raising barriers for smaller AI entrants. Surprising data shows AI boosts some creative outputs—studies indicate 20% productivity gains for writers—but displaces entry-level jobs. Trade-offs pit short-term creator protections against long-term societal benefits from accessible AI.