29 June 2025 by Mondo2.0

If I have an idea for a story and I submit it to ChatGPT, who is the author?

One of the most frequent questions I get about the new generation of AI services concerns copyright, in relation to what is generated through artificial intelligence.

If I have an idea for a story and I have ChatGPT (or Claude or Gemini or something) write it, who is the author?

The answer, based on the current context and the quality and type of services currently provided, is quite simple, while the implications that follow are particularly complex.

AI as a Tool: I am the author.

Artificial intelligence is a tool; I provide the initial idea, the creative direction, and I perform final editorial supervision, possibly requesting further changes. So, I’m the author.

Or at least that’s how it will be until a new generation of services, in a few decades, not soon, will have evolved and trained enough to create a story on their own, without any initial human input.

The Training Question: The Crux

But let’s stay with today, training is one of the most difficult knots to untie.

The idea is mine, but what if I use content created by others to train the AI? If I draw significantly from other sources, am I still the author?

The issue becomes further complicated when we consider that Large Language Models are trained on huge amounts of textual data, often without the explicit consent of the original authors.

This raises fundamental questions about the very nature of assisted creativity: can we consider original a work that emerges from a system that has “digested” millions of previous works?

Things get complex, very complex. It’s necessary to define the boundary between creativity, originality, and plagiarism on a case-by-case basis, based on the content used in the training (quantitatively and qualitatively), the input provided by the author, and the uniqueness/originality of the final content.

It’s highly likely that the use of AI services will increase the amount of content generated easily and immediately, content that draws excessive inspiration from others. This could require both a profound rethinking of the principle of originality and the development of AI-based tools capable of assessing a work’s level of creativity.

A Paradigmatic Example: The Transformation of the Classic

A concrete example can help clarify the complexity of the issue. With very little input, I asked Claude to revise the first canto of Dante’s Inferno.

Are we dealing with an original work, plagiarism, a tribute, an insult, or, more likely, just plain rubbish? Can I claim to be the author (even though, truth be told, I’m not particularly proud of it)?

The First Canto of the Cybernetic Nexus

In the midst of downloading my digital existence,

I found myself in a forest of crazy servers,

where the main pathway glitched into infinite loops.

Shit, describing this damned datacenter is hacker stuff,

so corrupted, full of viruses and malware,

which renews the blue screen of death in my processor!

It’s so badass that the album format is little more than that;

but to talk about the code I found hidden there,

I’ll tell you about the other bugs I scanned for you in the dark.

I don’t remember exactly how I entered this matrix,

I was so full of quantum lag at that moment

that the real operating system I abandoned amidst the crashes.

Just like banal posts, selfies, and ballets, will we also be inundated on the Internet with novels and works generated “one click”?

Towards a New Concept of Creativity

We’re subverting the concept of authorship, which doesn’t necessarily mean the author wrote the story himself. After all, we’ve long since ceased to cultivate our own land, nor are we directly responsible for building our cars on assembly lines. Consequently, we will be responsible for the initial design and supervision, not the actual production of our works.

What has already happened in many practical contexts is now also happening in creative contexts, with the automatic, surrogate generation of images, texts, videos, audio, …

All of this will quickly change our romantic notions of creativity, the role, status, and rights of the author; facts that will likely lead, in the next five years, to a significant regulatory overhaul of copyright, and the benefits and burdens that come with it.

In a highly dynamic generative context, copyright will take on a “liquid” form, changing depending on the context and, as mentioned above, will likely be judged by AI itself: a true paradox.

This evolution opens the way to Luciano Floridi’s definition of meta-author based on “… a model of authorship that is neither entirely human nor entirely artificial, but emerges from the interaction between human intention and agential capacity .” ( Distant Writing: Literary Production in the Age of Artificial Intelligence ).

What awaits us?

Generative artificial intelligence is democratizing access to creative tools, allowing anyone to produce content that previously required years of training. A phenomenon similar in terms of opportunity (and entropy) to what has already happened on the internet through social networks.

This phenomenon will lead to a polarization of the creative market: on the one hand, a sort of “premium segment” where purely human creativity will retain high value as a cultural luxury good; on the other, a mass segment dominated by low-quality, free AI content, not always generated from reliable sources.

New professions will emerge focused on curating published works and others on algorithmic creative direction and validation of generated content, while other traditional creative professions will risk obsolescence, overwhelmed by this World 2.0.

Mondoduepuntozero