News-Media Alliance Releases Whitepaper on AI, Copyright and the Fair Use Fallacy
Click play to LISTEN to the article below
In the rapidly evolving landscape of artificial intelligence (AI), publishers are finding themselves at a crossroads. The emergence of generative artificial intelligence (G-AI) systems has brought about a technological revolution, but with it comes a pressing question: How do we protect our creative content in this new digital age?
The Challenge: A recent white paper by the News-Media Alliance highlights a critical issue facing publishers today—the pervasive copying of expressive works by G-AI systems. These systems, which include large language models (LLMs) like ChatGPT, are trained on vast amounts of data, often without the consent or compensation of the original creators. This practice not only infringes on copyright laws but also threatens the very foundation of the publishing industry.
Understanding the Impact: The white paper underscores that G-AI systems are commercial products built on the backs of creative contributors. By using the creative output of publishers to train their models, G-AI developers are essentially bypassing the need for users to engage directly with the original sources. This diminishes web traffic and undermines the trust and brand value that publishers have cultivated over years.
Fair Use Fallacy
The argument that such use of copyrighted content is a fair use does not hold up under scrutiny. The white paper points out that the training of LLMs is not sufficiently transformative and is overwhelmingly commercial, which are key factors against the fair use defense. Moreover, the outputs of these models directly compete with the protected content, further impacting the market for the original works.
The Fair Use Fallacy in the AI Ecosystem: A Closer Look
In the burgeoning field of artificial intelligence, the use of copyrighted material to train generative AI systems has sparked a contentious debate over the boundaries of fair use. The News-Media Alliance (N-MA) has taken a firm stance on this issue, challenging the notion that such practices fall within the scope of fair use—a defense often cited by AI developers.
Understanding Fair Use: Fair use is a legal doctrine intended to balance the interests of copyright holders with the public’s right to use works in ways that contribute to education, commentary, criticism, or scholarship. It is not a blanket permission slip for all forms of use, especially when such use could harm the potential market for or value of the copyrighted work.
The Four Factors of Fair Use: The fair use defense is traditionally anchored in four factors:
- The Purpose and Character of the Use:
- AI developers argue that the use of copyrighted content is transformative. However, the N-MA asserts that training AI models is not transformative in nature. The purpose of these models is commercial, and the character of the use does not add new expression, meaning, or message to the original copyrighted works.
- The Nature of the Copyrighted Work:
- Copyrighted works of a factual nature are more likely to be considered fair use when used appropriately. However, news articles and other media content often contain creative and subjective elements, making them less likely to fall under fair use when used in AI training.
- The Amount and Substantiality of the Portion Used:
- AI systems require large volumes of data to learn and generate outputs. The substantiality of the content used from individual publishers can be significant, and in many cases, entire articles or works are copied verbatim. This extensive use exceeds what would typically be considered fair use.
- The Effect of the Use on the Potential Market:
- Perhaps the most critical factor is the impact on the market. The N-MA highlights that AI-generated content directly competes with the original works, potentially cannibalizing the market and reducing the incentive for publishers to invest in creating new content.
The Economic Argument: The N-MA emphasizes that the economic impact of AI on publishers is profound. By using copyrighted material to train AI without compensation, developers are benefiting from the investment of publishers in journalism and content creation without contributing to the costs. This undermines the economic viability of the publishing industry and threatens the livelihood of those who create content.
The Need for a New Understanding: The N-MA’s position is that the current interpretation of fair use by some AI developers is flawed and does not consider the full economic and creative impact on the publishing industry. They advocate for a reevaluation of fair use in the context of AI and a more balanced approach that recognizes the rights and contributions of all parties involved.
Conclusion: The fair use fallacy is a critical issue at the intersection of AI and copyright law. As AI continues to advance, it is imperative that the publishing industry, legal experts, and policymakers work together to ensure that fair use is applied in a manner that is just, equitable, and sustainable for the future of content creation.
A Call to Action
As publishers, it’s imperative to understand the legal and economic implications of G-AI systems. The white paper concludes with several actionable recommendations:
- Engage in Dialogue with AI Developers:
- Publishers should initiate conversations with G-AI developers to discuss the use of their content. This dialogue is crucial for establishing boundaries and understanding how publisher content is being utilized in AI training datasets.
- Advocate for Transparency:
- Demand that G-AI developers disclose the datasets used in training their models. Transparency is key to ensuring that the use of copyrighted material is acknowledged and compensated.
- Educate on Infringement:
- Work with industry groups to educate policymakers and the public about the nature of infringement in the context of AI. It’s important to clarify why the unauthorized use of content to train AI models is not fair use.
- Licensing Agreements:
- Publishers should push for the development of standardized licensing agreements that respect copyright laws and ensure fair compensation for the use of their content.
- Monitor AI Developments:
- Stay informed about the latest developments in AI technology to understand how it may impact the use of published content. This knowledge will be crucial in negotiating terms with AI developers.
- Legal Action:
- Consider legal avenues to protect intellectual property. The white paper suggests that current copyright laws support the rights of publishers, and legal action may be necessary to enforce these rights.
The Path Forward
- Collaborative Innovation:
- Publishers should seek partnerships with AI developers to create new products that leverage AI while also benefiting the original content creators. This could include joint ventures or collaborative platforms that utilize AI to enhance the reach and impact of published content.
- Develop New Revenue Streams:
- Explore how AI can be used to create new revenue streams. For example, publishers could license their archives for AI training in a controlled manner that respects their copyrights and generates income.
- Invest in Technology:
- Invest in AI and machine learning to understand its capabilities and limitations. This can help publishers to better negotiate with AI firms and to develop their own AI-driven tools and services.
- Policy Advocacy:
- Engage with lawmakers to shape policies that protect publishers in the AI ecosystem. This includes advocating for laws that recognize the value of creative content and ensure fair compensation for its use.
- Educational Initiatives:
- Create educational materials and workshops for staff and the wider publishing community on the implications of AI for copyright and content creation.
- Strategic Alliances:
- Form alliances with other publishers and content creators to have a stronger collective voice when dealing with AI developers and policymakers.
The white paper is not just a call to arms; it’s a blueprint for collaboration. Publishers are encouraged to engage with G-AI developers to create a fair ecosystem where innovation thrives without compromising the rights and revenues of content creators.
The AI revolution is here, and it’s reshaping the way we think about content creation and distribution. As publishers, we must be proactive in protecting our work. By understanding the legal landscape, advocating for fair compensation, and collaborating with technology developers, we can navigate the AI climate effectively and ethically.
Co-written with ChatGPT-4 | Image by Dall-E 3 and Adobe Firefly