Nigeria's Copyright Act 2022 and Generative AI: Addressing Fair Dealing Challenges
“People recognize intellectual property the same way they recognize real estate. People understand what property is. But it's a new kind of property, and so the understanding uses new control surfaces. It uses a new way of defining the property”.
- Michael Nesmith
Introduction
Generative AI is changing the way we create content, combining human creativity with machine learning to produce art, writing, music, and more. As this technology grows quickly, copyright laws are struggling to keep up. Nigeria’s recent Copyright Act of 2022 is an important step forward, but it still doesn’t answer many questions about AI-generated content. Can laws made before AI really handle a technology that’s changing the way we create?
One key issue here is fair dealing, an important rule in copyright law that lets people use small parts of copyrighted material without needing permission, as long as it’s for specific purposes. Traditionally, fair dealing has allowed uses like teaching, research, critique, and news reporting. But what happens when AI, which relies on huge amounts of data to improve, starts using copyrighted materials on a large scale?
In this post, we’ll look at how fair dealing works in the age of AI, the challenges it brings, and why Nigeria’s copyright law may need updates to handle this new reality.
What is Fair Dealing, and Why Does it Matter?
Fair dealing is a legal principle that allows limited use of copyrighted works without permission under certain circumstances. Traditionally, this principle supports activities like education, research, critique, and news reporting. For example, quoting a book in a review or copying a passage for personal study generally falls under fair dealing. Section 20 of the Copyright Act of 2022 outlines several key exceptions, or instances, where fair dealing can apply:
Private Use: You can make copies of copyrighted material for personal use, like transferring a legally purchased song onto your phone.
Parody, Satire, and Pastiche: This allows creators to remix existing works for humorous or transformative purposes. However, there’s a requirement that the work must add new meaning, not simply imitate the original.
Non-commercial Research and Private Study: Using copyrighted materials for non-commercial research is permitted, which supports academic work and personal learning.
Criticism, Review, or News Reporting: This exception allows the use of copyrighted works in critiques, reviews, or news reporting, though proper credit to the original source is required.
These fair dealing exceptions aim to strike a balance between protecting creators and enabling public access to knowledge and creativity.
When Does Something Count as Fair Dealing?
Determining whether a particular use qualifies as fair dealing requires Nigerian courts to consider various factors:
Purpose and Character of the Use Uses that contribute to public knowledge, such as educational, critical, or transformative purposes, are more likely to be deemed fair. For example, using a brief excerpt from a book in an educational presentation can be permissible if it serves an educational purpose without directly competing with the original.
Nature of the Copyrighted Work Works with factual content, like news articles, are often more accessible under fair dealing, as they provide public knowledge. In contrast, artistic works, which are seen as embodying more creative effort, receive greater protection.
Amount and Substantiality of the Portion Used Courts examine both the quantity of the work used and the importance of that portion to the original work. While using small excerpts may qualify as fair, taking an excerpt that is central or essential to the work could still infringe on copyright.
Effect on the Market If the use of copyrighted content undermines its market value or affects the creator’s income, it is less likely to be considered fair. Courts assess whether the new work competes with the original, potentially diverting sales or diminishing its market.
Challenges and Risks Associated with Fair Dealing in the Age of AI
The application of fair dealing to AI is complex, especially given how generative AI operates. Let’s look at some of the core legal concerns and challenges that arise when generative AI intersects with fair dealing:
1. Scale of Use vs. Limited Use
Fair dealing was designed for limited use cases. AI, however, needs massive amounts of data for training, often consuming entire libraries of text, art, or music to create new content. This scale goes far beyond what the fair dealing exceptions likely anticipated. Nigerian courts would need to examine whether this “wholesale” data ingestion truly aligns with the idea of fair use or if it crosses into outright copyright infringement. In particular, if the AI’s outputs directly compete with the original works, this could fall outside fair dealing’s intended boundaries.
2. Transformative Use and AI
Transformative use is a cornerstone of fair dealing. This means the new work should add new meaning, purpose, or value to the original, think of it like re-imagining a story rather than just reprinting it. The tricky part with AI is that while some models produce highly original output, others can generate work that closely mimics the original material used in training. Courts would need to ask: are these AI-generated works transformative enough, or are they effectively duplications? Deciding where to draw the line between innovation and imitation is a legal grey area that Nigeria’s Copyright Act doesn’t fully address.
3. Market Impact of AI-Generated Content
Fair dealing assumes that the new use of a work won’t harm its market value. But what if AI-generated content saturates the market, competing with or even replacing original works? For instance, if an AI model generates images that rival a photographer’s work, it could impact their ability to earn a living. This potential harm to the market goes against the spirit of fair dealing, which aims to protect creators’ livelihoods.
Risks for AI Developers
Without clear guidelines on fair dealing, AI developers face considerable legal risks:
Copyright Infringement: Training AI on copyrighted data without permission could spell trouble. For commercial AI applications, the stakes are especially high. Without a license or some fair dealing protections, AI companies could find themselves in hot water.
Liability for Users’ Actions: Developers might also be held liable if users create infringing content using their AI tools. Courts may assess whether the AI’s design encourages copyright violations, further burdening developers.
Commercialization Challenges: Licensing and selling AI models trained on copyrighted data can be legally risky. Without proper licensing, the commercial use of these models could lead to legal disputes.
Data Privacy Concerns: AI developers must also consider data privacy regulations. Using personal data in training without consent may lead to privacy violations, with potential penalties that impact both reputation and revenue.
Risks for Content Creators
Content creators, too, face uncertainties as AI reshapes the creative landscape:
Loss of Control: Creators may lose control over how their work is used in AI training. When AI models pull from copyrighted materials, creators have little say in how their work is repurposed, potentially without credit or compensation.
Opaque Use of Data: AI training datasets are often hidden from public view, making it challenging for creators to know if their works are included. This lack of transparency complicates efforts to protect intellectual property rights.
A Call for Reforms: Updating Copyright Law for AI
As Nigeria looks to future-proof its copyright framework, here are some proposed changes to balance innovation with creator rights:
For AI Developers
Refining Fair Dealing: Introducing specific fair dealing exceptions for AI training could provide clear guidelines and reduce legal uncertainty.
Developing a Licensing Framework: Establishing a licensing system for AI training data could ensure fair compensation for creators while giving AI developers a legal route to train their models.
Clarifying Transformative Use: The definition of transformative use could be updated to clarify how it applies to AI. This would help courts distinguish between truly innovative outputs and those that simply copy existing content.
Safe Harbour Provisions: Introducing safe harbour protections for developers who follow strict guidelines could encourage responsible AI innovation while respecting copyright law.
For Content Creators
Improving Transparency: Clear rules on dataset transparency could help creators understand how their work is used in AI training, fostering greater trust.
Revenue-Sharing Models: Setting up revenue-sharing agreements could ensure that creators benefit financially when their work is incorporated into AI models.
Conclusion
The rise of generative AI is a game-changer for copyright law in Nigeria. To ensure both technological growth and the rights of creators are protected, we need to rethink the concept of fair dealing. The Copyright Act 2022 is an important first step, but we must make more changes to keep up with the rapid developments in AI. By reassessing fair dealing, improving licensing rules, and increasing transparency, Nigeria can create a space where technology and creativity can grow together successfully.