Why Generating Art from AI Models is Not Theft: A Comparison of Derivative Works versus Transformative Works

Lately I have been observing the debate surrounding the generation of art  through AI models and whether it constitutes theft. While some may argue that it is, I strongly believe that it is not theft, but rather a form of transformative works.

To understand why I believe this, it is important to define the two concepts. Derivative works are works that are based on an existing work, while transformative works are works that use an existing work to create something new and original.

In the case of AI generated art, the original work is the data set that is used to train the model. The model then generates new works based on that data set, creating something that is original and unique. This is a clear example of a transformative work.

Furthermore, it is important to consider the role of the AI model in the creation process. AI models do not have the capacity for personal expression or intention, and therefore, the works generated by the model cannot be attributed to a human author. This is in contrast to derivative works, where the author of the derivative work is human and the work is attributed to them.

It is also worth noting that copyright law has evolved to encompass new forms of creative expression and technology, including computer-generated works. This evolution has helped to ensure that artists and creators are protected, while also promoting creativity and innovation.

Take this as just my opinion, but I believe that generating art through AI models is not theft, but rather a form of transformative works. The works generated by the model are original, unique, and cannot be attributed to a human author. Additionally, copyright law has evolved to encompass new forms of creative expression and technology, ensuring that artists and creators are protected.

Back to blog