The AI Code of Ethics for Artists
A rough guide for artists navigating use cases for artificial intelligence.
What is this?
This AI Code of Ethics is an attempt at a guide to help artists understand their role in the spreading impact that AI is having on the arts community and to provide rough guidelines on how to behave in relation to AI in their creative life. The ethics are meant to be understood as a whole, as they all relate to each other.
Who is this for?
This code of ethics is meant for creative professionals and amateurs. However, people who are concerned about their AI use might benefit from these guidelines.
Why write this?
When talking about AI with friends or artistic peers, the conversation often gets to a point of someone saying, "Well hopefully someday some government body will regulate this stuff." That hasn't happened, so it seems prudent to develop some cultural guidelines to help guide behaviour in the meantime.
How were the ethics chosen and created?
I have reviewed other excellent examples of codes of ethics, consulted with creative professionals, workshopped ideas with friends, brainstormed, attended lectures, participated in discussion boards, and worried about the four ethics. This is the result of that effort!
Is this a binding or comprehensive document?
Absolutely not. There are likely many holes left open by these ethics and every use case of AI is different. For example, much could be said on the difference between generative and other types of AI. This document is meant to be a tool for how to think about this issue, not a book of laws.
There may be some individual circumstances that prevent a person from following one of these guidelines. These circumstances are personal and must be addressed with honesty from the individual level.
There may be some individual circumstances that prevent a person from following one of these guidelines. These circumstances are personal and must be addressed with honesty from the individual level.
How can I help?
Please consider adding your name to the list at the bottom of this page, and share this page and these ideas with your community.
Ethic 1:
Avoid using AI to replace "Core Work" in creation or consumption.
Explanation
"Core Work" is the principal work that is being done by an artist. Painters paint, trombonists play trombone, dancers dance, and writers write. All creatives wear many different hats, but most "big C" Creativity is being done by artists working in their core capacity, and can often be identified as creative decision making with regards to process, content, execution, and more. By following this ethic, we can preserve much creative work from being replaced.
Examples
Against This Ethic
- A composer uses AI to write the melody for their next symphony. The composer's "Core Work" is writing music, and they have outsourced that to an AI.
- A video game designer uses AI to replace a voice actor for a character. The video game designer isn't replacing their own "Core Work", but they are using an AI to replace the voice actor's "Core Work".
- A singer-songwriter uses an AI tool to make album art for their next release. This is directly replacing the "Core Work" of a graphic designer.
- A University Student uses an app that generates lo-fi beats for studying. This directly replaces the "Core Work" of artists doing similar work.
- A photographer uses an AI tool to highlight a subject in their photograph in order to perform adjustments using a mask. This work used be done manually, but this tool makes that work that they would already do themselves much faster.
- A painter uses an AI tool to summarise their sales data from a spreadsheet and highlight trends. This is not the painter's "Core Work", and isn't creative in nature.
- A YouTuber uses an AI tool to add captions to their video and checks the results for correctness. This work used be done manually, but this tool makes that work that they would already do themselves much faster.
Grey Areas and discussion
This is perhaps the most controversial, most difficult to follow, and most important of the four ethics here. It is tempting to use AI for many applications, including the core work of an artist.
One of the most common expressions that I've heard about AI is "It's going to be able to replace artists!" This is correct. Generative AI will someday soon be able to replicate most creative work. This is something that we should be trying to address on a cultural level.
There is a case to be made for a fifth ethic: "Don't use AI to mimic people". This ethic does cover that idea, as a person is by themselves their "Core Work".
Grey areas for this ethic are plentiful. The example about the painter using AI to find trends in their sales data shows work that could feasibly be done by a business coach or consultant. This is where individual circumstances have a great impact on how this code of ethics is used. One should take extreme care when using a budget to "excuse" using AI to do work that a human could do, as it is the most common explanation used, and very difficult to find a limit to this reasoning. Following this ethic may, and probably should, be costly and/or painful for many individuals and organisations.
Another grey area that has already been explored is using AI in the creation of art ABOUT AI. Each individual artist will have to decide how they proceed with this type of content, although it is fairly reasonable to speculate that this topic will become less common as the excitement of generative AI as a whole becomes less prominent.
There is much to consider about the homogenising of human expression. This is mostly addressed in Ethic #4.
One of the most common expressions that I've heard about AI is "It's going to be able to replace artists!" This is correct. Generative AI will someday soon be able to replicate most creative work. This is something that we should be trying to address on a cultural level.
There is a case to be made for a fifth ethic: "Don't use AI to mimic people". This ethic does cover that idea, as a person is by themselves their "Core Work".
Grey areas for this ethic are plentiful. The example about the painter using AI to find trends in their sales data shows work that could feasibly be done by a business coach or consultant. This is where individual circumstances have a great impact on how this code of ethics is used. One should take extreme care when using a budget to "excuse" using AI to do work that a human could do, as it is the most common explanation used, and very difficult to find a limit to this reasoning. Following this ethic may, and probably should, be costly and/or painful for many individuals and organisations.
Another grey area that has already been explored is using AI in the creation of art ABOUT AI. Each individual artist will have to decide how they proceed with this type of content, although it is fairly reasonable to speculate that this topic will become less common as the excitement of generative AI as a whole becomes less prominent.
There is much to consider about the homogenising of human expression. This is mostly addressed in Ethic #4.
Ethic 2:
Disclose any use of AI in work that can be seen by other people.
Explanation
People should have a right to consent to consuming content that has been created or significantly assisted by AI. Information is the driver of consent. Artists using AI in a context that will be seen by a consumer must reveal to that consumer that the AI is being used, and in what way.
Examples
In Line with this Ethic:
- A video game designer uses a AI tool to add captions to a voice actor's performance. They add a line to the game credits saying that the captions were built with AI.
- A costume designer uses AI generation to duplicate sketches in different colors for a client. They tell the client that the additional sketches were generated with AI.
- A YouTuber uses an AI tool to break a large video into smaller chunks. They publish the chunks with a line in the description that says they've used an AI tool to do this part of the editing.
Grey Areas and discussion
The major issue in this ethic is when and how to disclose to a consumer when AI is being used. AI is used often, and to write large disclaimers or content warnings in front of a piece of artwork will likely become cumbersome for everyone involved.
As the internet becomes different to navigate and as AI curation (see Ethic 3) becomes more common, it may be possible to automatically filter AI content from one's life using tagging. This is effective for social media, but other measures will have to be developed and used for print media, billboards, and public music.
Photography competitions are already practicing this ethic by separating individual competition categories between AI assisted, AI generated, and AI-less photography. This is a positive institutional step.
As the internet becomes different to navigate and as AI curation (see Ethic 3) becomes more common, it may be possible to automatically filter AI content from one's life using tagging. This is effective for social media, but other measures will have to be developed and used for print media, billboards, and public music.
Photography competitions are already practicing this ethic by separating individual competition categories between AI assisted, AI generated, and AI-less photography. This is a positive institutional step.
Ethic 3:
Avoid using AI as a curator or judge of artwork.
Explanation
Humans are the target of artwork. Therefore, humans should be the sole judge and curator of that artwork. AI should never serve as a formal adjudicator or curator for any competition, exhibition, or job opportunity.
Examples
Against This Ethic:
- A creative writing competition has received 500 entries and has asked an AI database tool to cull the entries to reach 25 finalist documents. This would be using AI to judge human artwork as unworthy (or worthy).
- A hiring committee for a performance position uses an AI to look for keywords in candidates' CVs to remove weaker candidates. This allows any bias of the AI's programming to judge human candidates.
- A singer-songwriter asks an AI to choose the best title for their new album. This would be using AI to judge one idea as better than another.
Grey Areas and discussion
There are many reasons that this ethic is required and this document couldn't possibly cover them all. Philosophically, humans should not be comfortable with using machines to judge their artistic output on principle, but there are many practical reasons that this should be avoided. For example, the ACLU reported as early as 2018 that an automated hiring tool that showed clear bias against women candidates for internal hiring. There have been demonstrated reports of racism and other undesirable outcomes from AI. When judging humans, having significant impact on their financial, artistic, and creative futures, it is best to leave this work to a human. Although the systems we have for adjudication and selection are not perfect, they will not be improved by AI.
One grey area for this Ethic comes in the form of selection algorithms for social media and streaming services. This technology has already become common and replaces the radio for many music listeners. There is significant discourse on how streaming and social media has impacted artists that should be used to inform consumer decisions about how to manage selection algorithms. Someone who wanted to completely eliminate AI from their artistic consumption would have to consider purchasing a lot of music and subscribing to individual artists.
One grey area for this Ethic comes in the form of selection algorithms for social media and streaming services. This technology has already become common and replaces the radio for many music listeners. There is significant discourse on how streaming and social media has impacted artists that should be used to inform consumer decisions about how to manage selection algorithms. Someone who wanted to completely eliminate AI from their artistic consumption would have to consider purchasing a lot of music and subscribing to individual artists.
Ethic 4:
Avoid using AI to replace meaningful human connection.
Explanation
The homogenisation of human communication and connection should be feared, not celebrated. Human uniqueness and imperfection should be preserved, and AI should not be used to replace it. Artists should strive to maximise their human connection.
Examples
Against this Ethic:
- A professor encounters writer's block when preparing a presentation. They have a short discussion with an AI to resolve the trouble when, a few years ago, they would have spoken to a grad student or a colleague. The grad student or colleague has missed an opportunity to communicate with a prominent member of their research community.
- A musician uses AI to write a difficult email to their bandmates about the group's future. The homogenising nature of AI writing prevents the musician's true feelings from being communicated, and the band feels alienated and distant.
- A University professor uses AI to help write their application for conference travel funding. The professor has written many similar applications before and there is a style guide on how to write the project description.
- An arts administrator uses AI to help write a new policy regarding ticket refunds. The AI tool helps them be clear in their new policy.
- A painter uses AI to write an email to a brush company to request a refund for defective products. The AI tool helps them reduce the mental load of writing this email and a human connection isn't nessesarily important in this exchange.
Grey Areas and discussion
Human connection is sometimes one of the most anxiety-provoking and difficult tasks that artists have to face, and it can be terribly tempting to ease the burden with having AI write emails or replace other communication structures. The magic of artists in their development of new words, phrases, and means of expressing scenarios cannot be replicated by generative AI, and we should lend every opportunity for those organic ideas to emerge.
This is perhaps the most grey of all the ethics, and will only become more so as the administrative demands upon artists are increased. This is when referencing Ethic #2 might be valuable when choosing to write an email, and some artists do this very well.
As technology and human life become more intertwined, simple mindfulness and intention may be the best tool to prevent the homogenisation of human communication.
This is perhaps the most grey of all the ethics, and will only become more so as the administrative demands upon artists are increased. This is when referencing Ethic #2 might be valuable when choosing to write an email, and some artists do this very well.
As technology and human life become more intertwined, simple mindfulness and intention may be the best tool to prevent the homogenisation of human communication.
Please consider adding your name to the list of artists endorsing this code of ethics.
- Rachel Smith: Musician - Teacher - Performer
- Victor Cabezas
- Nicholas Matherne: Artistic Director, Victorian Youth Symphony Orchestra
- Bon K
- Chelsea McBride
- Brian KM: Performer-Composer, Artistic Director Melbourne Rainbow Band
- Tracy Chan
- Kalia Page: Graduate Student, MM Instrumental Conducting, Temple University
- Kristen Dye: Band Director: Acton-Boxborough SD
- Lee Bradshaw: Composer
- Isaac Shieh: Natural Horn Virtuoso, New Music Specialist, Collaborator
I manage this list manually. If you want to be removed from this list, please let me know.
If you want to stay in touch with me, consider signing up for my email list.
If you want to stay in touch with me, consider signing up for my email list.