Generative Artificial Intelligence

Intellectual Property

Categories of Intellectual Property

There are five major categories of intellectual property:

1. Patents - Protection for inventions
2. Trademarks - Protection for products and their branding
3. Industrial Design - Protection for the visual design of objects 
4. Copyright - Protection for creative expressions such as literary works, art or photographs
5. Trade Secrets - Protection of proprietary information like recipes, formulas, or any kind of business data that could provide an advantage to a company

   ---

The intellectual property category that is the most prevalent in relation to GenAI is copyright. The two major questions concerning intellectual property and Generative AI are:

1. Were copyright laws violated during pre-training?
   1. Was data collected without the consent of the creator
   2. Was copyrighted material used to train the model
2. Who owns the content generated by Generative AI tools?
   1. The individual who created the prompt?
   2. The owner of the GenAI tool?
   3. Can generated content be protected by copyright?

Copyright violations during development

• To develop a GenAI model, an extensive dataset is essential. For large language models, this means an enormous amount of text, while image generators require a large collection of images, and so on. The source of this data, however, remains somewhat ambiguous. Many companies will not disclose their datasets or how they were created, leaving us guessing about the specifics of the data that has been used to train these models.
• This lack of transparency is an issue because it appears that few creators have been consulted about the inclusion of their creative works in these datasets. If copyrighted materials such as books, artwork, and videos have been incorporated into a dataset without the creators’ consent, it raises the question: has there been an infringement on their intellectual property rights?

Who owns Generative AI output?

• As of right now, copyright can only be granted to human creators, and GenAI tools are not human.  So, who owns the image created in Dall-E or the poem written by Copilot when those tools are prompted to write a haiku about a rainbow, and can those works be registered for copyright protections?  It depends.
• In March 2023, the United States Copyright Office published Copyright Registration Guidance: Works Containing Material Generated by Artificial Intelligence. This document outlines the requirements for a piece of work to be registered for a copyright.  The determining factor in whether or not a work which involved the use of GenAI is how much of the work was done by a human.  The example given in the guidance by the Copyright Office is:
  • "...if a user instructs a text-generating technology to “write a poem about copyright law in the style of William Shakespeare,” she can expect the system to generate text that is recognizable as a poem, mentions copyright, and resembles Shakespeare’s style.²⁹ But the technology will decide the rhyming pattern, the words in each line, and the structure of the text. ³⁰ When an AI technology determines the expressive elements of its output, the generated material is not the product of human authorship.³¹ As a result, that material is not protected by copyright and must be disclaimed in a registration application. ³²
• In other cases, however, a work containing AI-generated material will also contain sufficient human authorship to support a copyright claim. For example, a human may select or arrange AI-generated material in a sufficiently creative way that “the resulting work as a whole constitutes an original work of authorship.” 33 Or an artist may modify material originally generated by AI technology to such a degree that the modifications meet the standard for copyright protection. 34 In these cases, copyright will only protect the human-authored aspects of the work, which are “independent of ” and do “not affect” the copyright status of the AI-generated material itself."35 
• Please see the linked work for sources from the footnotes.