The rapid development of generative AI has sparked intense debate over how, or even if, creators should be compensated when their copyrighted works are used to train commercial AI systems. This issue pits the drive for technological innovation against the fundamental rights of authors to benefit from their creations, leading to diverse proposals for legal and economic frameworks that seek to strike a fair balance. The following three presentations from the Global Expert Network on Copyright User Rights Symposium in June 2025 explore this complex landscape from distinct legal, philosophical, and geopolitical perspectives.
The Geneva Centre on Knowledge Governance and the Program on Information Justice and Intellectual Property bring you three contributions to the AI Remuneration Debate.
PART 1: Christophe Geiger approaches the problem from a human rights perspective, arguing for a balance between the right to develop AI for cultural and scientific progress and the author’s right to benefit from their work. He critiques current systems, noting the “all-or-nothing” nature of the US “fair use” doctrine and the EU’s “bizarre” opt-out rule for text and data mining, which he believes fails to secure fair compensation for authors due to unequal bargaining power with publishers and producers. His central proposal is to replace the EU’s opt-out system with a mandatory statutory remuneration scheme for the commercial use of works in AI training. Drawing on the success of similar “remunerated exceptions” in Europe, which generate significant revenue, Geiger proposes that income from this scheme be distributed directly to creators. Geiger contends this model would uphold authors’ human right to fair remuneration without stifling innovation.
PART 2: Zachary Cooper reframes the debate by arguing that traditional copyright concepts are becoming obsolete in an age of infinite digital remixing and AI-driven content creation. He contends that focusing on authorship thresholds is futile because the line between human and machine creation is hopelessly blurred and impossible to audit reliably. Methods like watermarking are technically weak and easily circumvented. For Cooper, the real issue is the massive scale of AI generation, which makes copyright enforcement impractical and weakens creators’ negotiating power. He describes copyright as “a dam in an infinite river,” an outdated barrier against a constant flow of transformation. Instead of rigid ownership rules, Cooper suggests the future lies in collective licensing models and a greater emphasis on attribution and visibility, which would allow creators to capture value as their work spreads across massive platforms.
PART 3: Vitor Ido situates the remuneration debate within the political and economic context of Brazil and Latin America, presenting it as a crucial tool for regulating corporate power and protecting national creative industries. He explains that for GRULAC (Group of Latin American and Caribbean Countries), the issue is not just about copyright but about challenging the dominance of large, foreign-based platforms that exploit local content with little to no payment to creators. The discussion also encompasses cultural sovereignty, such as protecting the dubbing industry from AI-generated voices, and safeguarding the traditional knowledge of Indigenous communities from misappropriation. Ido highlights Brazil’s draft AI Bill, which proposes an inverse of the EU’s system: a mandatory remuneration right that includes a reciprocity clause and ties the payment amount to the size of the AI company, directly targeting the market power of major corporations. This approach frames remuneration as a strategic element in a broader agenda of economic justice and cultural preservation in the Global South.
