The intersection of artificial intelligence (AI) and copyright is rapidly becoming one of the most debated legal and policy areas globally. Governments and institutions are grappling with complex questions: Should copyright subsist in AI-generated content? How should AI training on copyrighted works be regulated? What are the economic implications of AI for copyright industries?
Recently, three reports have been published addressing these critical issues:
- Canada’s “What We Heard” Report on Copyright in the Age of Generative AI – A consultation summary detailing perspectives from Canadian stakeholders on copyright challenges posed by AI.
-
OECD Report on Intellectual Property Issues in AI Trained on Scraped Data – A deep dive into the legal, technical, and policy implications of AI models trained on vast amounts of scraped internet data.
-
U.S. Report on the Economic Implications of AI for Copyright Policy – An economic analysis from the U.S. Copyright Office on how AI might reshape copyright policy, industry incentives, and creative markets.
Below, I summarize the three reports.
Canada’s Consultation on Copyright in the Age of Generative AI
The Government of Canada conducted a broad consultation between October 2023 and January 2024 to explore the impact of generative AI on copyright law. This initiative was a response to growing concerns from cultural industries, artists, and AI developers about how copyright rules should adapt to emerging AI technologies. The resulting report, What We Heard, summarizes stakeholder feedback on key copyright issues.
The consultation focused on three main areas:
Text and Data Mining (TDM) and AI Training
The most contentious issue in the consultation was whether AI developers should be allowed to train their models on copyrighted works without permission.
• Creators and rights holders argued that AI training on copyrighted works without consent or compensation violates their rights. They emphasized the need for licensing frameworks that ensure fair remuneration.
• Technology companies and certain other user groups contended that text and data mining (TDM) does not infringe copyright because it extracts factual patterns rather than expressive content. Some called for explicit legal exceptions for TDM, similar to those in the EU and Japan.
• Many stakeholders supported greater transparency regarding what data AI systems use for training.
Authorship and Ownership of AI-Generated Content
Another major debate centered on whether AI-generated content should be eligible for copyright protection.
• Most participants favored maintaining human authorship as a requirement for copyright.
• Some stakeholders suggested AI-generated content could be protected under copyright if a human made substantial contributions.
• Concerns were raised about how AI might dilute creative markets by producing content that competes with human creators.
Infringement and Liability for AI-Generated Content
Stakeholders were divided on whether existing copyright infringement laws adequately address AI-generated works that closely resemble copyrighted material.
• Some advocated for clearer liability rules, particularly regarding who should be responsible when AI-generated content infringes copyrights.
• Others argued that existing copyright doctrines, such as substantial similarity tests, could sufficiently address these concerns.
• Transparency mechanisms, such as AI-generated content labeling, received some support.
The consultation also highlighted several adjacent concerns:
• Indigenous communities raised issues about the unauthorized use of traditional knowledge and cultural expressions in AI systems.
• Performers and public figures expressed concerns about the use of their likenesses in deepfake AI applications.
• Some stakeholders worried about AI’s impact on job security and unfair competition in creative industries.
The report did not attempt to outline how the Government plans to act in relation to the copyright challenges. The closest indication of the Government’s thinking was this paragraph in the conclusion of the report:
The government continues to consider how Canadian concerns posed by generative AI, including those raised by cultural and technology industries, might be addressed. In considering the policy issues at the intersection of copyright and AI, the government is particularly mindful of preserving incentives to create and distribute works provided in the copyright framework, while supporting Canada’s innovation strategy. The government recognizes that these policy issues have the potential to question the foundations of the copyright framework, but believes that copyright remains relevant in the age of generative AI. Over time, copyright law has always been resilient in adapting to technological disruptions—from the printing press to the advent of the Internet—and to changing business models in the marketplace.
OECD Report: Intellectual Property Issues in AI Trained on Scraped Data
The OECD’s Intellectual Property Issues in AI Trained on Scraped Data report examines how AI models are trained using data scraped from the internet and the resulting intellectual property (IP) challenges. The key these are as follows.
Data Scraping and Copyright
• Many AI models rely on massive datasets scraped from online sources, including copyrighted materials, raising concerns about potential infringement.
• Legal frameworks differ across jurisdictions, leading to uncertainty about whether AI training on scraped data constitutes copyright infringement, and if so under which laws
Policy Approaches
• The report suggests possible policy responses, such as a voluntary code of conduct for AI developers to ensure fair use of copyrighted works.
• It also explores technical tools, such as metadata tagging and licensing mechanisms, to allow rights holders to control how their content is used in AI training.
Legal and Economic Implications
• Increasing litigation over AI scraping practices, particularly in the U.S., highlights the need for clearer rules.
• The OECD suggests that standardized licensing agreements could help balance innovation and copyright protection.
The OECD report highlights the growing tension between AI innovation and copyright enforcement. It emphasizes the need for international cooperation on AI copyright policies, given the global nature of data scraping.
U.S. Report on the Economic Implications of AI for Copyright Policy
The U.S. Copyright Office’s report, Identifying the Economic Implications of Artificial Intelligence for Copyright Policy, explores how AI is reshaping the economics of creative industries and copyright law. It focused on the following:
AI-Generated Works and Market Competition
• AI-generated content could displace demand for human-created works, raising concerns about income loss for artists.
• Some argue that AI lowers production costs, making creative industries more accessible, but others worry about the devaluation of human creativity.
Copyrightability of AI Outputs
• Should AI-generated works receive copyright protection? The report examines economic incentives for human vs. AI creation and the potential need for a new copyright framework.
AI’s Impact on Copyright Holders
• The ingestion of copyrighted materials for AI training could affect incentives for creators, depending on how revenues are redistributed.
• The report explores whether compulsory licensing or revenue-sharing models could ensure fair compensation for copyright owners.
The U.S. report presents an economic framework for evaluating AI’s effects on copyright policy. While it does not prescribe specific solutions, it highlights the economic trade-offs involved in regulating AI-generated content.
The Road Ahead for AI and Copyright
These three reports reflect a rapidly shifting legal and economic landscape for AI and copyright. Key themes emerge across all three studies:
• AI transparency is a growing demand, with calls for clearer disclosure of training data sources.
• The role of copyright law in AI training remains contentious, with creators pushing for licensing and AI developers advocating for broader exceptions.
• Economic concerns about AI’s impact on creative industries are rising, with policymakers considering mechanisms to balance innovation and fair compensation.
With ongoing legal battles and policy discussions worldwide, the debate over AI and copyright is far from settled. However, these reports provide useful information for future regulatory decisions. As governments weigh their options, expect further legal challenges, legislative proposals, and calls for international cooperation in this evolving area of law, at least outside of the United States.
