Is ‘Fair Use’ Enough to Protect AI in the Global Courts?

In light of the constant advancements in the realm of Artificial Intelligence and the related technologies, it has become increasingly pertinent to establish and ensure that such advancements do not infringe Copyright and the related Intellectual Property rights. However, considering the lack thereof, of structured and established regulatory measures pertaining to Artificial Intelligence and related technology, numerous legal suits have been filed by content creators across the globe against the Artificial Intelligence developers. 

Deliberation on the Question of Copyright Infringement and the Defence of ‘Fair Use’ Doctrine by Artificial Intelligence technology in the First Indian AI Case

An ongoing, yet pivotal copyright dispute in India, i.e., Ani Media Pvt Ltd Vs. Open Ai Inc & Anr. wherein ANI Media has initiated legal proceedings against OpenAI, alleging the unauthorized utilization of its journalistic content for training the ChatGPT language model. ANI in its lawsuit has contended that the appropriation of its data by Chat GPT infringes upon copyright statutes and has inflicted considerable reputational damage, particularly due to the dissemination of fabricated narratives erroneously attributed to its agency. In defense, OpenAI has asserted that it did not directly access ANI's paywalled content but instead used publicly available information to train its model. The company invoked the ‘Fair Use’ doctrine, which essentially, permits the use of copyrighted material without permission for purposes such as commentary, criticism, or educational use. OpenAI vide its submissions has maintained that its use of publicly available data aligns with the exception under ‘Fair Use’, as the data is not being directly copied but is only utilized to train its Artificial Intelligence model to generate novel and unique content. 

This case underscores significant legal questions regarding the application of Copyright law in the context of generative Artificial Intelligence and the prospective judicial interpretation of the legal issue at hand. The concept of ‘Fair Use’ is central to this debate, as it will determine whether OpenAI's training practices shall be considered permissible and in accordance with the provisions of Copyright law. At this juncture, it is important to understand that the Doctrine of ‘Fair Use’ has been a key defense for tech companies in the United States of America, and although it is a given, that the application and interpretation may vary across different jurisdictions, the present suit in India shall essentially prompt valuable guidelines, precedent and discussions on how other countries will address similar cases, given the lack of precedent and legislative regulations governing Artificial Intelligence and data usage in the realm of Copyright law.

These legal challenges due to the lack of legislative and judicial guidelines, highlight the complexities of using proprietary material in Artificial Intelligence training. Training Artificial Intelligence models on copyrighted or proprietary data without proper authorization can lead to infringement claims, as seen in the lawsuits against OpenAI and Stability AI. Additionally, there are concerns about the potential exposure of confidential information. If sensitive or proprietary data is entered into a generative Artificial Intelligence tool, it may inadvertently incorporate this information into its training, leading to unintended disclosures or misuse. Therefore, it is important to draw inference from the interpretation of the said legal complexity in foreign jurisdictions. This article has limited the scope of comparative discussion to the American jurisdiction considering the stark difference in the interpretation of the ‘Fair Use’ doctrine by the Indian and American legislative and judicial bodies.

Interpretation of Fair-Use Doctrine in the United States of America

Within the United States, the ‘Fair Use’ doctrine has been pivotal in the defense strategies of numerous technology corporations against allegations of copyright infringement. The prevailing argument posits that the employment of copyrighted materials for Artificial Intelligence training without direct reproduction constitutes a "transformative" use, thereby qualifying for protection under Fair Use statutes. However, this defense remains subject to judicial scrutiny, as evidenced by ongoing litigation where artists, authors, and content creators have expressed apprehensions about Artificial Intelligence’s capacity to replicate their works without appropriate authorization or remuneration. These concerns are central to several lawsuits challenging Artificial Intelligence companies' practices. 

• The Authors Guild vs. Google Inc.: In this case, the Authors Guild sued Google for scanning books and making their content searchable via the Google Books project. The court ruled in favor of Google, stating that the use of copyrighted books was transformative and fell under ‘Fair Use’. This case has set a significant precedent for Artificial Intelligence companies, as it helped establish arguments about "transformative use" in the context of Artificial Intelligence models training on vast datasets. 

• Case Involving Visual Artists: In the case of ‘Andersen v. Stability AI Ltd’, a group of visual artists sued OpenAI for using their artwork to train the company's generative Artificial Intelligence model, without permission. The artists argue that their copyrighted works were used to teach Artificial Intelligence model how to create new images that mimic their style, which they consider an infringement of their intellectual property rights. The case is still ongoing and reflects how artists are increasingly concerned that generative Artificial Intelligence might undermine their livelihoods and the value of their work. 

• The New York Times Case: The New York Times has sued OpenAI and Microsoft for the unpermitted use of Times articles to train GPT large language models. The case could have a significant impact on the relationship between generative Artificial Intelligence and Copyright law, particularly with respect to ’Fair Use’, and could ultimately determine whether and how Artificial Intelligence models are built. OpenAI has used copyrighted New York Time’s content by arguing that it falls under the doctrine of transformative use. OpenAI has also claimed that its Artificial Intelligence models generate new, unique content rather than simply reproducing or copying NYT articles, emphasizing that the purpose of its models is to create transformative, generative outputs. OpenAI also acknowledges some instances of exact reproductions, which it attributes to bugs that they intend to fix, stressing that the purpose of the Artificial Intelligence is not to replicate articles but to create new works based on the data it has absorbed. However, the New York Time’s case against OpenAI is strong, as they argue that OpenAI’s use of their content is commercial, potentially harms their market, and involves the substantial use of their copyrighted material. The court will need to balance the interests of fostering innovation in Artificial Intelligence development with the rights of content creators, particularly in light of concerns over the commercial nature of the use and its potential impact on the New York Times's market. The outcome is uncertain, as the case will involve applying ‘Fair Use’ principles to a novel context involving generative Artificial Intelligence.

In view of the above, it is pertinent to note that in the ANI Media lawsuit, OpenAI is expected to defend its actions under the ‘Fair Use’ doctrine, arguing that training its model on publicly available data is a form of transformation. OpenAI contention that it does not "copy" the content but rather uses it to generate new and unique outputs, could set a significant precedent for Artificial Intelligence companies in the United States of America., as well as in India, potentially allowing it to use copyrighted materials for model training, provided they can demonstrate transformative use. However, the interpretation and the scope of applicability of the same, may vary in both jurisdictions owing to the differential approach adopted by the judiciary in India and the United States of America pertaining to the doctrine of ‘Fair Use’.

How Can Artificial Intelligence Developers Avoid Legal Liability and Content Creators Monitor IP Infringement

To navigate this evolving landscape of Artificial Intelligence and its alleged infringing technology, specialized IP counsel plays a pivotal role in guiding technology companies through the complexities of Artificial Intelligence-related legal issues such as:

Developers: Ensuring Compliance and Transparency

Artificial Intelligence developers must adhere to legal standards concerning the acquisition and use of data for model training. This includes complying with regulations on licensing and compensating IP owners for data incorporated into training datasets. While the current responsibility partly falls on Artificial Intelligence tool users to verify proper licensing, emerging legal precedents are likely to shift more responsibility onto developers. To mitigate risks, developers should prioritize transparency by maintaining detailed records of AI-generated content, including development platforms, settings, metadata, and tags. Such audit trails can protect against IP infringement claims and demonstrate that outputs were not intended to copy or infringe upon existing works.

Creators: Proactive Protection and Collaboration

Individual content creators and brands should actively safeguard their IP portfolios by monitoring digital and social channels for derivative works and utilizing search tools to examine large-scale datasets. For brands with valuable trademarks, monitoring should extend to the stylistic elements of derivative works, as stylistic similarities may indicate misuse of brand content. IP legal counsel can assist in regular landscape monitoring to help creators protect their proprietary assets. Additionally, creators should establish frameworks for working with AI-generated material to defend against claims of infringement, plagiarism, or misappropriation. Understanding what constitutes ‘Fair Use’ or sufficiently "transformative" work is crucial in the realm of AI-generated content. Properly establishing ownership over AI-assisted creations is essential, especially if creators intend to profit from such works. Collaborating with IP counsel can help navigate these areas and ensure compliance with evolving legal standards.

Businesses: Evaluating Contracts and Ensuring Compliance

Companies should regularly assess contractual language and terms to ensure clear addressing of IP ownership issues related to Artificial Intelligence and transparent guidelines for Artificial Intelligence use in innovation. This includes confirming proper licensure of training data from Artificial Intelligence platforms and securing broad indemnification for potential IP infringement. Vendor and customer agreements should incorporate AI-related clauses to protect IP rights and support the registration of authorship and ownership of AI-generated works. As the legal landscape evolves, keeping legal counsel informed about generative Artificial Intelligence usage is essential. Organizations should consider creating checklists for contract modifications to mitigate unintended risks associated with Artificial Intelligence use. By taking these proactive steps, companies can navigate the complexities of Artificial Intelligence and Intellectual Property law, ensuring that their innovations are legally protected and ethically developed.

Conclusion

In summary, as Artificial Intelligence technologies continue to advance, the role of specialized IP counsel becomes increasingly vital in guiding developers, creators, and businesses through the evolving legal landscape. By adopting proactive measures and collaborating with legal experts, stakeholders can navigate the complexities of Artificial Intelligence and Intellectual Property law, ensuring that their innovations are legally protected and ethically developed.

Thus, it is evident that the intersection of Artificial Intelligence development and Copyright law is a dynamic and multifaceted arena, with each jurisdiction adopting distinct approaches that reflect its legal traditions, economic priorities, and cultural values. The ongoing discourse and litigation in this domain underscore the necessity for a nuanced understanding of both technological advancements and the legal structures that govern them.

In view of the same, the ongoing legal dispute as discussed in this article between ANI Media and OpenAI has significant implications for the future of Artificial Intelligence development and the protection of intellectual property rights. 

The resolution of this case could set a precedent influencing Artificial Intelligence companies' data collection practices and their obligations to content creators. A ruling in favour of content creators may compel Artificial Intelligence companies to obtain explicit permissions or pay licensing fees to use copyrighted materials, potentially altering current Artificial Intelligence training methodologies. Conversely, a decision favouring Artificial Intelligence companies might allow greater freedom in training models, potentially accelerating the development of generative technologies but also raising concerns about fair use and creator compensation. The outcomes of such legal disputes will play a crucial role in shaping the future of both Artificial Intelligence technology and Copyright law worldwide.

References:

1. CS(COMM) 1028/2024.

2.  Mira T. Sundara Rajan (UC Davis Law School), Is Generative AI Fair Use of Copyright Works? NYT v. OpenAI, February 29, 2024. Accessible at:  

3. https://copyrightblog.kluweriplaw.com/2024/02/29/is-generative-ai-fair-use-of-copyright-works-nyt-v-opena/

4. No. 13-4829 (2d Cir. 2015).

5. 700 F. Supp. 3d 853 (N.D. Cal. 2023).

6. Supra 2.

7.  Audrey Pope, NYT v. OpenAI: The Times’s About-Face. Accessible at:  

8. https://harvardlawreview.org/blog/2024/04/nyt-v-openai-the-timess-about-face/.

9.  Michael Dilworth, AI & Intellectual Property: Artificial Intelligence Legal Implications. Accessible at: https://www.dilworthip.com/resources/news/artificial-intelligence-and-intellectual-property-legal-issues/.

10. AI and Copyright Infringement: Creator’s Guide. Accessible at: 

11. https://www.scoredetect.com/blog/posts/ai-and-copyright-infringement-creators-guide?.