Is training AI to generate content a violation of copyright laws?
This article explores the intricate relationship between generative AI and copyright law, investigating the potential infringement issues arising from training models using copyrighted material. With the proliferation of AI technologies like CHATGPT and DALL-E, concerns about the utilization of protected works for training purposes have surfaced. The article discusses landmark judgments worldwide, emphasizing the need for clear guidelines to balance innovation and copyright protection. Factors such as the fair use doctrine, consent, and licensing are analysed to determine the ethical use of generative AI models.
INTRODUCTION
As the field of Artificial Intelligence rapidly advances in this globalizing world, important questions arise about the existence of copyright law. How are AI outputs regulated in the same way as infringement aspects for training models that use the works of the original author? Copyright law is based on the originality of the work, but AI is a combination of human intelligence and machines that can produce an output from the infringed work of authors. The application of AI is vast, and training it on a large scale can lead to copyright infringement. However, as the use of AI is on a massive level, maintaining fair use of copyrighted work is becoming difficult. Therefore, various cases and issues regarding copyright infringement are arising while training generative AI.
This paper aims to discuss how training generative AI leads to copyright infringement determined on a case-to-case basis of various landmark judgments around the world. Further, it will analyse how AI and copyright are interconnected and what further steps can be incorporated to stop the training of generative Artificial Intelligence from infringing copyrighted works. A Strict guideline needs to be developed by the Copyright offices around the world so that prominent companies around the global south and north can implement these guidelines effectively while training AI based on the data collected by the copyright owners.
HOW AI AND COPYRIGHT ARE INTERLINKED?
The development of Artificial Intelligence (AI) technology has brought about a new era of innovation and machine learning. However, the use of copyrighted material in training AI models has raised concerns and legal difficulties worldwide. This has led to lawsuits being filed, as many believe that Generative AI models like CHATGPT and MIDJOURNEY are using the works of protected authors and artists to create derivative works.
As a result, there is no consensus on whether this falls under the Fair Use principle under copyright law. Each case is determined by the courts on an individual basis. If it is ruled that copyrighted material cannot be used to train AI models, it will have a significant impact on the training process of these models.
For copyright protection, the work must meet the criteria of originality. The question is whether AI possesses originality, as the training of AI relies on existing data and algorithms created by humans. Therefore, the primary question is whether training generative AI models leads to an infringement of copyright.
For instance, the recent development of DALL-E, a generative AI model that creates images from textual descriptions, used approximately 650 million images of publicly available online licensed resources for its training purpose. This could be seen as a case of infringement when copyrighted work is used without prior permission and royalty fees. Therefore, it is essential to establish a standard way to determine whether AI models' use of copyrighted materials falls under Fair Use or not.
IS TRAINING GENERATIVE AI INFRINGEMENT OF COPYRIGHT?
With the rapid advancements in technology, particularly in the field of generative AI, it has become increasingly challenging to differentiate between original human-created work and AI-generated work. This poses significant complexities and challenges concerning intellectual property laws, including the ambiguity surrounding ownership, infringement, and ethical concerns related to the utilization of unlicensed data to train generative AI models.
To address these issues, it is essential to consider a few preconditions, including fair use doctrine, consent of the original owner, and authorized licensing, while determining whether training generative AI will result in copyright infringement. Fair use is a fundamental principle in modern societies, which ensures that information in the public domain can be used for the greater good of society and to create new designs and statements based on it. However, the usage should be reasonable, justifiable, and comply with the four criteria for fair use.
The four criteria that must be examined before invoking the fair use principle are as follows:
1. The nature and intent of the use: When utilizing copyrighted material, it is important to keep in mind that it should only be done for educational, business, or research purposes. It is crucial to consider the nature and intent of use, ensuring that it is reasonable and justifiable. By doing so, you can avoid any potential legal issues and ensure that you are using the material in a fair and ethical manner.
2. The composition of the protected work used: It is important to note that any creative works used in a project or presentation should be based on factual or non-fictional material only. It is imperative that such usage should not be based on any fictional works, as it may lead to misinterpretation or confusion among the audience. Therefore, it is recommended to exercise caution when selecting creative works to ensure that they are accurate, informative, and relevant to the topic at hand.
3. The proportionate size and significance of the work used: When using a copyrighted work, it's important to make sure that the amount used is appropriate compared to the entire work. In other words, the portion used should not be so significant that it can be considered the "heart" of the work. This helps to ensure that the original creator's rights are respected while still allowing for fair use in certain circumstances.
4. The measure of the impact on the potential market and value of the original copyrighted work: When using a copyrighted work, it is crucial to ensure that it does not negatively affect the market value of the original work. This is the most significant factor in determining whether the usage of the copyrighted material is fair or not. Therefore, careful consideration must be given to this criterion to avoid any adverse impact on the potential market and value of the original work.
By adhering to these preconditions and criteria, we can ensure that the usage of generative AI models is ethical and complies with copyright laws.
CONCLUSION:
The intersection of generative AI and copyright law raises complex questions. As AI technology advances, the potential for copyright infringement becomes a significant concern. The legal landscape surrounding this issue is evolving, making it imperative to establish clear guidelines and standards. To address these challenges, it is crucial to consider factors such as the fair use doctrine, consent of the original owner, and authorized licensing. Moving forward, there is a pressing need for international cooperation and the development of strict guidelines to regulate the training of generative AI models while fostering innovation and creativity in AI development and ensuring the protection of intellectual property rights.
