Introduction

The idea of technological advancements outpacing legal protections is a modern phenomenon that scholars are seeing across different fields. Digital privacy has become a widely discussed topic recently that has sparked debate over what protections are needed or not, and which areas should be considered for additional legal protections. As artificial intelligence continues to grow, society’s dependence increases, and legal issues continue to arise. This chapter explores the growing gap between technological innovation and legal protections in the modern United States. The impacts of this discrepancy have been large across the board, and solutions to this issue are critical to maintain order in the technological sector of our society.

During the 2020s, artificial intelligence underwent a meteoric rise in development, notably with models like OpenAI’s ChatGPT, Anthropic’s Claude, and Google’s Gemini. These models, built with a combination of generative AI and large language models, are nearly unrecognizable compared to what came before. Based on massive sets of training data, they can output human-level writing with supported research, generate images and video, and scan the internet for information at a rate incomparable to humans (Foote). These models have raised issues as well, notably in the world of copyright. To collect the training data required for these models to learn and develop, companies feed copyrighted writing, art, videos, and sounds into their models (Buick). Additionally, there are questions about who owns the outputs created by these models. This chapter aims to describe the current landscape regarding AI and copyrighted materials.

Connection to STS

The intersection between science, technology, and society is complex and applies to a vast number of theories and ideas we deal with regularly. As artificial intelligence grows, societal dependence grows as well, creating increased awareness of these platforms. In turn, developers continue to adapt and adjust as needed, further advancing the technology. Furthermore, those who create these new technologies typically do so in response to a societal or scientific need. These interactions are what continue to develop our society, and our legal system intertwines to protect these ideas and the people.

There are many questions that emerge when considering this overlap. For example, when technology outpaces legal protections, we must identify who is responsible for regulating it and how to protect those who are using it. Additionally, it is important to note that studies have shown that this gap between technology and the law disproportionately affects minority groups. According to “Gender Shades: Intersectional Accuracy Disparities in Commercial Gender Classification”, facial recognition software tends to have higher error rates for minority groups as opposed to Caucasian groups (Buolamwini and Gebru). While examining these questions, we have to remember the connection to science, technology,y and society to properly evaluate the impacts and contributions from all groups.

Copyright battles between creators and AI companies depict an active debate between the worlds of science, technology, and society. People on all sides of this issue have their own opinions on the rights maintained by copyright holders and whether or not AI companies should have the ability to feed any work of art into their models during development. While people can debate the ethics amongst themselves, what will ultimately determine the relationship between AI companies and copyright holders will be determined in a legal setting.

Legislation has already been set in action to address the issue, such as the Generative AI Copyright Disclosure Act of 2024. This bill, still in development, would introduce a precedent where AI companies must fully disclose all of the training data they use with full transparency (Bondari). Additionally, multiple lawsuits have already been filed alleging the illegal use of copyrighted works as training data, notably Andersen v. Stability AI and The New York Times Company v. Microsoft Corporation et al. Each of these lawsuits consists of copyright holders (artists in Andersen, and The New York Times newspaper) who are alleging improper use of their work as training data (Bondari). While both of these cases are ongoing, the rulings will likely introduce valuable precedent in the future of AI and copyright.

Topic Information

In regards to training data, most AI companies operate under the doctrine of fair use, which was established by the Copyright Act of 1976. Fair use permits creative works to be utilized without direct permission from the copyright holders in specific situations based on four factors. These are: the purpose for use, the nature of the copyrighted work, how much of the work is taken, and whether or not the use is considered “transformative”, and considering if the new work will replace the original (Blaszczyk). These guidelines provide the precise debate points between both sides of the argument.

In the case of Andersen v. Stability AI, the plaintiffs focus on factors one, three, and four. The company’s training data, which consisted of a massive 600 million image set, consisted of numerous copyrighted works that were fed to Stable Diffusion, their AI image generation model. The plaintiff’s claim that the use of their works is both nontransformative and for a commercial use that could replace their work (Samuelson). Essentially, they argue that the collection of their works to be a template for AI image generation is derivative and poses an economic threat to their well-being. As this case advances and is eventually settled, a precedent for image use in AI training data could be set, which may either protect or deny the ability of artists, photographers, and filmmakers to prevent their copyrighted works from being used in AI training data.

Another likely landmark case to be litigated in the near future is The New York Times Company v. Microsoft Corporation et al. Rather than being based on image training data, The New York Times is alleging that both Microsoft and OpenAI have committed copyright infringement by feeding NYT articles into training data for their generative AI models ChatGPT and Gemini. The NYT claims that since they were not consulted for their works nor were they compensated for their use, reparations are due (Jurcys). The NYT also includes some unique claims in their lawsuit, claiming that the outputs of these models would often consist of fully verbatim rewrites of its articles, and even instances with incorrect information included. With this claim, the NYT argues that they are owed more than just monetary compensation, and also hopes to establish a precedent wherein Microsoft and OpenAI can no longer infringe on copyrighted materials for training purposes (Jurcys). This would be the landmark change introduced if set in stone, which would set a precedent for future copyright holder powers.

While these two cases are just some of the many active litigations against AI companies regarding copyrighted works being used as training data, the eventual decisions made in court on all of these cases will set a clearer precedent on what permissions these companies have pertaining to fair use, copyright, and training data.

When it comes to generative AI outputs, questions have arisen over the ownership of text, video, image, and other creations. Can a person enter a prompt, and then take the output and file a copyright over what has been generated? In the United States, the answer is no. The U.S. Copyright Office does not grant copyright to works that lack human authorship (Kugler).

This is similar to a case from the 2010s for the ownership of selfies taken by monkeys. In 2008, a nature photographer named David Slater set up cameras in Indonesia near some macaques. One of the macaques ended up taking several images, some of which became very famous “selfies” that were shared online. Eventually, the images began to be posted online by people other than Slater, who began to challenge this, claiming he had the copyright as he set up the conditions for the photos to be taken. However, in court, it was determined that the photos lacked a human author and therefore could not be filed under copyright (Bavitz). This ruling has persisted into AI output arguments, where AI-created works cannot be protected by copyright law (Kugler).

In countries other than the United States, including the United Kingdom, Ireland, South Africa, and others, there are some laws that permit some protection over some AI-created works (Kugler). Essentially, it isn’t impossible across the globe to protect some AI outputs, but it is very limited. Perhaps in the future, new judicial actions will be taken to provide copyright benefits to the AI companies or users. However, until changes are made, AI outputs are considered generally uncopyrighted.

When technology continues to advance over time, the need for legal frameworks and protections continues to grow with it. In our modern society, technological dependence is heavy, and we typically complete daily tasks that are reliant on some form of cell phone or computer. For example, almost all assignments in universities across the country are completed and submitted online instead of on paper. With this growing reliance, the law struggles to keep up for a multitude of reasons:

1. Stare Decisis and Precedent: Law relies on previous cases, making it difficult to establish new, updated reform.
2. Timing: Creating new law takes time, regardless of the method (Legislation or Case law)

While there are thousands of new technologies emerging each year, one sector that has struggled to adapt is digital privacy. One common issue is surveillance, which has been heavily litigated at both the state and federal levels. According to “Taxonomy of Privacy” by Daniel Solove, surveillance can lead to self-censorship and is often difficult to appropriately regulate by law. With this in mind, we have to remember that there are large amounts of data privacy laws already in place; however, as our technologies evolve and require more personal data to operate, we must adapt our legal protections.

Another sector that is battling this discrepancy is the Artificial Intelligence platforms being rolled out. Companies all across the world are beginning to integrate these systems into typical decision-making procedures to be more efficient; however, there are significant legal risks in doing so. As discussed in the American Bar Association’s article, “Top Six AI Legal Issues and Concerns for Legal Practitioners”, the increased usage of AI in corporations can lead to output bias, employee discrimination, job displacement, and ethical risks.

These issues across different technological products create several consequences for users and society as a whole. One of the biggest concerns of advocates for better legal reform is the exploitation of users of these technologies. Typically, companies require personal data to operate their technologies, and if not legally regulated appropriately, this data can fall into the wrong hands. Additionally, lacking the appropriate legal framework to regulate these platforms can create confusion for companies, leaving room for major ethical violations.

Scholars and legal professionals have long debated the appropriate reform to combat this issue. Three general ideas have been agreed on, but the efforts are ongoing. Many figures agree that creating laws at the statutory level is an appropriate measure for regulating the usage of emerging technologies and protecting consumers. Within this, there would need to be adequate ethical guidelines in place to monitor companies from acting irresponsibly with the data collected. Creating a solution to this issue is complex and difficult to navigate, but it is important to maintain legal order in America. Authors of “The Growing Gap between Emerging Technologies and Legal-Ethical Oversight” write, “If we allow that lag time [Technology outpacing the law] to increase, it will grow exponentially until both ethics and law will be realistically viewed as an irrelevant antique of a time long past” (Marchant).

Conclusion

The introduction of new technologies tends to advance the world overall and ease processes, but the gap between these technological advances and the law presents a pressing issue for modern society. As we continue to develop, the need for reform in the legal field grows simultaneously. We must find a way for the law to evolve as efficiently and consistently as our technology.

More specifically, as AI technologies continue to develop and become more prevalent in STEM fields, it is important to understand the legal arguments surrounding its input and output; The field of biochemistry is no exception to these interactions. When a user asks ChatGPT a question or guidelines for designing an experiment, where is the information coming from? Giving proper credit and ensuring the main idea of these articles is understood by AI platforms is vital, so understanding where these models get their information is incredibly important. Additionally, AI could be used in experimental design. If a person uses these outputs and reports them in their work, do they own those procedures? Do they need to report this use? While this applies to almost every scientific field, developing a concrete relationship between copyright and generative AI is an important future development that will define the use of AI across all fields and what is accepted regarding both inputs and outputs.

From an STS perspective, this issue displays the connection with both society and technology, emphasizing the importance of creating measures to regulate both effectively. Overall, closing the gap between technological advancements and legal protections will take time and collaboration from different industries to protect consumers and corporations.

References

Amdur, E. (n.d.). 6 critical – and urgent – ethics issues with ai. https://www.forbes.com/sites/eliamdur/2024/01/24/6-critical–and-urgent–ethics-issues-with-ai/

Bavitz, C., & Albert, K. (2024, May 30). The “monkey selfie” case: Can non-humans hold copyrights? Berkman Klein Center. https://cyber.harvard.edu/events/2018/luncheon/01/monkeyselfie

Blaszczyk, M., Stanley, K. D., & McGovern, G. (2024, November 20). Artificial Intelligence Impacts on Copyright Law. RAND. https://www.rand.org/pubs/perspectives/PEA3243-1.html

Bondari, N. (2025, February 4). Ai, copyright, and the law: The ongoing battle over intellectual property rights. IP Technology Law Society. https://sites.usc.edu/iptls/2025/02/04/ai-copyright-and-the-law-the-ongoing-battle-over-intellectual-property-rights/

Buick, A. (2024). Copyright and AI training data—transparency to the rescue? Journal of Intellectual Property Law and Practice, 20(3), 182–192. https://doi.org/10.1093/jiplp/jpae102

Buolamwini, J., & Gebru, T. (2018). Gender Shades: Intersectional Accuracy Disparities in Commercial Gender Classification. https://doi.org/https://proceedings.mlr.press/v81/buolamwini18a/buolamwini18a.pdf

Foote, K. D. (2026, March 18). A brief history of generative AI. Dataversity. https://www.dataversity.net/articles/a-brief-history-of-generative-ai/#the-2020s-and-smarter-chatbots

Harvard Business School. (n.d.). 5 ethical considerations of AI in business – HBS online. https://online.hbs.edu/blog/post/ethical-considerations-of-ai

Jurcys, P., & Fenwick, M. (2024). NY Times vs Microsoft and openai: Should it be an “easy” fair use case to decide? SSRN Electronic Journal. https://doi.org/10.2139/ssrn.4685275

Kugler, L. (2024, October 4). Who owns AI’s output? – communications of the ACM. Who Owns AI’s Output? https://cacm.acm.org/news/who-owns-ais-output/

Marchant, G. E., & Allenby, B. R. (2013). The growing gap between emerging technologies and legal-ethical oversight: The pacing problem. Springer.

McNeely Law Firm. (2025, September 29). The legal implications of AI in business decisions making. McNeelyLaw LLP. https://www.mcneelylaw.com/the-legal-implications-of-ai-in-business-decisions-making/

MyCase. (n.d.). Top six AI legal issues and concerns for Legal Practitioners. American Bar Association. https://www.americanbar.org/groups/law_practice/resources/law-technology-today/2025/ai-legal-issues-and-concerns-for-legal-practitioners/

Samuelson, P. (2023). Generative AI meets copyright. Science, 381(6654), 158–161. https://doi.org/10.1126/science.adi0656

Solove, D. J. (2006). A Taxonomy of Privacy. University of Pennsylvania Law Review, 154(3), 477–560. https://doi.org/https://scholarship.law.gwu.edu/cgi/viewcontent.cgi?params=/context/faculty_publications/article/2074/&path_info=A_Taxonomy_of_Privacy.pdf