A great aspect of using any types of neural networks for image generation is that anybody can use them without any pre-existing technical understanding and even if they have never tried creating artwork before. Even the best image generation model is designed for accessibility, meaning everybody can create unique graphics simply by entering a text prompt.
The potential debate emerges when we think about how DALL-E is trained as this involves feeding the algorithm huge datasets drawing on art styles and pieces that have been created by professional artists.
Where an artist has not consented to their work being used in AI model training or where a user instructs the AI to generate a graphic mimicking the style and technique of an artist, the thorny question of copyright becomes even more complex.
Is it Possible to Register Copyright Against an AI-Generated Artwork?
Generally, copyright bodies do not accept copyright applications for graphics developed using algorithmic processes based on the thinking that an AI artwork cannot be 100% owned by the user. Part of that rationale is that the AI has done most of the work and that the art produced is a culmination of external sources the model draws on to interpret a text prompt.
However, that also depends on the generator you are using and for what purpose. Some platforms state that all creations made using AI are copyright free, whether in a legal, personal or commercial application.
Copyright rules also vary in different jurisdictions. For example, the US authorities state that copyright authorship is only allocated to people who have created an artwork as a human and can prove their work complies with originality standards.
New standards and rules are also evolving in other countries where copyright can be granted, provided the creator can show a minimum level of human involvement in the artwork design and end product.
Why Is AI Artwork the Topic of So Many Copyright Debates?
Regardless of the fact that in most places, you cannot register copyright against an AI graphic, the potential for copyright infringements continues to be fiercely challenged. Lawsuits have been raised, citing that licensed and commercially owned images have been used to create AI graphics that deliberately copy existing works.
Even though AI artwork is rarely copyrighted, the source material or images input to build the final graphic could be. Best practice use of AI can easily avoid the possibility of legal conflict by avoiding a text prompt that explicitly states a style, brand, artist or photographer to incorporate into the image.
Creators, artists, and photographers often either register copyright against their original works or issue licences for authorised use. They can also apply a Creative Commons licence to prevent works from being intentionally copied–this wouldn’t stop a person from trying to mimic their work or style, but it means they have recourse to take action if this happens without their approval.
How Will Copyright Standards Change for AI Art in the Future?
As AI-generated art becomes more widely used and better understood, it is extremely likely that attitudes and copyright rules will evolve, particularly when more and more companies want to use AI artwork in ad campaigns and other commercial projects.
Countries with stricter image copyright rules may take longer to adapt or to create frameworks to protect the creative materials of both professional artists and AI creators. It also isn’t probable that any significant change will be introduced in the near future, given the complications around enforcing copyright protections against an AI algorithm.
In the meantime, it is wise to be aware of the debate and to steer clear of generating images that are obviously based directly on an artist’s work to prevent any potential exposure to copyright issues.