Carrie Bickner

Understanding Responsibility for Human/AI-Created Objects

Editorial Note: This document has been updated and incorporated into the MARTI Framework master document, which provides a comprehensive overview of the latest developments. For the most current and complete information about MARTI, please visit: https://carriebickner.com/marti/

Who is responsible for an object created by human/generative AI? An introduction to our first piece of documentation:

Marti Contribution Table and Schema Document (Version 1.0)

The best way to begin to address this problem is by creating a record of who created what and to describe generative AI involvement in a way that is specific and flexible.

This question gets at the heart of our concerns about the authenticity of things that are created by AI. The proverbial question goes something like this; is the photograph I am looking at real?

People have been able to manipulate images for some time. Even before image editors like PhotoShop came along, humans could “airbrush” photos to make us look better, or to place ourselves in funny or compromising settings. Our inability to grasp the difference between a “doctored” photo and the unaltered original seems quaint now, given the potential corruptibility of images, data and any other object in the digital space. It is even possible that a generative AI could produce real-world objects with dubious provenance.

The best way to begin to address this problem is by creating a record of who created what and to desctibe generative AI involvement in a way that is specitic and flexibe.

It is not going to be enough to distinguish AI-generated objects from human-generated ones. We are in a hybrid world where collaboration between generative AI and humans is, or will be soon, de rigueur.

Main Entry

If I were to have my ChatGPT edit this file, which I will not (I need to explain why and how I co-created the MARTI Framework, and to be able to discuss the documentation fluently), ideally I could express that process with metadata. The same would be true for an image that I create, co-create, edit, modify or otherwise adulterate with generative AI.

We need systems that track the nature of these exchanges and collaborations, and we need to begin with a statement of intellectual responsibility.  In library cataloging terms, we refer to this statement as the “main entry,” and a catalogers first job is always to establish this datapoint.

People, corporations, and each instance of any given AI will be captured and expressed in the MARTI Framework.

Other Forms of Contribution and Collaboration

These entities will have roles beyond authorship.  The MARTI Framework will make use of controlled lists of human and machine generative entities, and will be able to use these lists in any number of rolls.  MARTI will be able to express agency in a way that is extensible, flexible and adapts to the variety of iterative collaborations we expect to see between generative AI and people. This will allow us to capture and express the role of artists who coax new works out of generative AI, prompt engineers who generate marketing material, or programmers who train LLMs. Each form of creation and collaboration will be captured, recorded and made discoverable using the MARTI Framework.

Harmonizing with Copyright Law

In the US, it is not possible to copyright material generated by AI. Intiutivly, this makes sense. Why give rights to an agent who can’t benefit from holding a copyright?  I am sure that there are many other reasons for this prohibition.

In the world of AI generated content, objects, materials, it will rarely be accurate to as the product of an AI and only an AI. There will almost always be a human element in whatever AI creates. I hope that by introducing a more nuanced understanding, and record, of intellectual responsibility, we can more easily incorporate authorship statements for generative AI/human collaborations, and that this will help us harmonize these new practices with the laws of the land.

I am aware that my view is shaped by the fact that I live in the United States and welcome input from all fields and in other parts of the World.