Generative AI: Is Your Copyright Safe?

May 3, 2024

Generative AI is a branch of artificial intelligence that creates or “generates” entirely new content and is revolutionizing creative fields. From music composition to graphic design to authorship and more, these tools have the potential to democratize artistic expression and accelerate innovation – providing better access to creative license and information than ever before. However, an issue remains with this exciting technology: copyright infringement. With the growing prevalence of AI and its integration into most industries, an important conversation is developing amongst users and policymakers as to the legislation needed to ensure copyright protection.  

How Does Generative AI Work? 

The internet hosts a vast collection of copyrighted materials. Millions of books, news articles, images, artwork, videos, and personal data can be found at the click of a button from anywhere in the world. Generative AI tools are trained on this data, as the nature of computer and machine learning is based on accessing information and using it to recognize patterns and relationships. When prompted, generative AI has the power to produce entirely new outputs based on what it has learned, such as a song in the style of a favorite musician, a poem inspired by a classic novel, or a photorealistic image of a mythical creature. This data can come from a variety of sources such as open-source content, which includes websites, articles, and code freely available on the internet, or copyrighted information, which includes text, images, music, and ideas protected by copyright law. 

The Copyright Quandary 

While using open-source content to train generative AI is somewhat unproblematic, incorporating copyrighted information into this data collection process raises concerns. Copyright law allows for some limited, transformative use of copyrighted material, however, the line between "fair use" and “infringement” is often blurry. For example, training a generative AI model on vast amounts of copyrighted data might be considered infringement, or if the AI generates derivative works that are too similar to the originals. 

AI researchers argue that generative models do not simply copy protected content. Instead, they learn underlying patterns and concepts, allowing them to create genuinely new works without infringement. But an increase in copyright lawsuits against AI image-generation platforms suggests otherwise, as some argue that AI models can "memorize" copyrighted elements, leading to outputs that infringe on the original work in ways that may not be immediately recognizable. 

Copyright infringement of generative AI is seen most recently in a developing lawsuit set in motion by several companies against Microsoft. According to CNBC, “newspaper publishers in California, Colorado, Illinois, Florida, Minnesota and New York said that Microsoft and OpenAI used millions of their articles without payment or permission to develop artificial intelligence models for ChatGPT and other products,” indicating a larger issue not yet addressed with the data collection, merging, and output methodologies used in generative AI. 

The Need for Improved Policies in AI 

Generative AI offers immense potential, but the copyright infringement issue requires ongoing dialogue and potential legislative changes. Global government agencies are struggling to catch up in creating the policies needed to address these issues inherent and specific to the internet world – making this a critical area of industry focus. One necessary change would be to adapt current copyright laws to provide clearer guidelines and define what exactly is acceptable training data, as well as offer stricter copyright protections and punishments for unauthorized use. 

Generative AI is a powerful tool with the potential to reshape creative industries. However, the copyright question remains a hurdle. Finding a balance between fostering innovation and protecting intellectual property will be key to unlocking the full potential of this exciting technology. As discussions continue and legal frameworks evolve, the expectation for a future where generative AI thrives hand-in-hand with respect for creative ownership grows more promising.  

The Future of AI Education 

As generative AI tools become more accessible, it is crucial to educate users on copyright limitations. Understanding responsible use and respect of intellectual property rights is essential for fostering a thriving generative AI ecosystem. Capitol Technology University offers a variety of bachelor’s, master’s, and doctoral programs focused on AI education and the ethical usage of this innovative technology. Students can also join our new AI Center of Excellence (AICE), which is dedicated to advancing the frontiers of artificial intelligence through innovative research, education, and industry collaboration. 


Considering a career in AI? Learn more by contacting our admissions office.