Copyright Infringement in the Age of AI

Sep 15, 2024

2 min read

AI is a creative; capable of generating works of fiction, composing music, structuring quality news articles and more.

As technology evolves, so too does the creative process of machines. Through neural networks (building blocks of AI that mimic human neurons) machines have developed the ability to create works of art in a similar fashion to humans. Machine learning algorithms learn through the data input they receive, whether the system uses supervised learning where-in the programmer is specifically programming it or through unsupervised learning, where-in the system learns through un-labelled data, the data remains the source of the learning. This data contains human-created content, millions of it. Therefore, regardless of the way in which a machine learning algorithm operates i.e., supervised, unsupervised or reinforcement learning, it learns from humans and does not credit an individual creator but rather a collection of data which may or may not be labelled.

"Although data itself cannot be copyrighted, you may be able to own a copyright in the compilation of the data. Creative arrangement, annotation, or selection of data can be protected by copyright." - Kent State University. The laws surrounding data protection vary depending on an individual country's laws. The emergence of AI coincides with the emergence of cyber crimes including copyright infringement in the digital era. It is for this reason that 17 authors including George RR Martin, Jodi Picoult and John Grisham have filed a lawsuit against OpenAI because of its ChatGPT. ChatGPT is an open-source large language model that produces "human-like" text based on a dataset that has been available on the internet up until 2021. That dataset has produced a training ground for the system which has become the most popular open-source large language model, with over 100 million global users as of February 2023. The authors accuse OpenAI (the parent company of ChatGPT) of “systematic theft on a mass scale," and are suing the company for an undisclosed sum of money. In response to the class-action suit, an OpenAI spokesperson stated the company is, "having productive conversations with many creators around the world, including the Authors Guild”. This lawsuit opens up a larger conversation about theft of material that may or may not be copy-written, including music, visual art and more. AI learns from humans and without the proper attribution, AI steals from humans.

AI is not a creative, AI is a system, a model, built to enhance our work not replicate it. Without the proper constraints and laws to govern AI "creativity" and large language models in particular, writers are at risk. The Picoults and Grishams have power, they have access, resources and voices. However, they do not speak for all writers. Writers do not need to sell millions of books for their work to matter, they do not need to be published in the Wall Street Journal for their words to have currency and value. It is more difficult for an unknown writer to go up against OpenAI, but as long as their words line the corridors of the internet, their words pave a path for AI. Their words need to be recognised and their voices need to be heard, AI cannot and must not be allowed to silence the writer

Sources and links:

https://libguides.library.kent.edu/data-management/copyright#:~:text=Although%20data%20itself%20cannot%20be,can%20be%20protected%20by%20copyright.