Music and technology, working together in harmony, can open up new exciting creative horizons. Independents embrace technological progress and have continually adapted to new commercial opportunities and innovations.
As a global creative community that wants to foster responsible Artificial Intelligence development in and with music, we are keen to engage with AI developers who prioritize human creativity, safety, transparency and respect for copyright.
WIN has collectively produced the following principles as a guide for our community, as well as for policymakers and AI developers around the world. These principles apply to any creative content used at any stage of AI development, including any used to train models that generate synthetic output.
This is not intended to be an exhaustive list, but rather is provided as a starting point to enable open and collaborative engagement.
AI development is subject to copyright
- The use of music and other related creative content (such as lyrics, audio-visual material and images) in any stage of AI development (including “training” and outputs) is subject to copyright. No exceptions should be introduced.
- Laws relating to AI should be developed or strengthened to protect music to a high standard, with cross border consistency.
- All uses of music and related creative content and metadata, as well as creators’ voice, image, name and likeness, require explicit permission in advance of “training”.
- Creators are entitled to fair and appropriate remuneration for such uses on fairly negotiated terms.
Prioritizing a human-centered approach
- Creativity is inherently human and is at the heart of all great music and other art.
- GenAI should be developed and applied to assist human artistry and creativity rather than replace it.
- Facilitating human creativity must remain central to GenAI development, regulation and legal frameworks.
- Content generated by AI without human creativity cannot qualify for copyright.
Safety of creators, fans, consumers and the public
- Tackling unauthorized deepfakes and other misappropriation of identity is a priority to ensure public safety.
- Any content generated with full or partial AI input needs clear labeling to promote public trust and informed consumer choice.
- Personal data of all people, including creators, should be protected from unauthorized use in GenAI models.
- Creators’ voice, image, name and likeness should be protected from unauthorized input or output use in GenAI models.
Transparency as a fundamental element
- Transparency and accountability must be built in at all stages of GenAI development and use.
- Record keeping, reporting and auditability of all data used to develop and operate GenAI systems must be legal obligations.
- AI-generated content must be clearly labeled and/or watermarked.
- Credit and attribution must be given when creators’ work is used.
Ethical AI development hand-in-hand with music
- Our community supports innovation and encourages responsible AI development.
- AI development must avoid bias and promote opportunities for all.
- AI systems in music should foster diversity of creators, genres and markets.
- We welcome AI developers who proactively consult and work in harmony with the international music community.
Music is central to humanity and plays a fundamental role for wellbeing in society. The global independent community engages with different forms of AI and welcomes the tremendous possibilities they offer to assist us to create, produce, distribute, and promote music.
We invite all AI developers to work hand-in-hand with us to help level the playing field for all, explore licensing and commercial opportunities together, and create a fully functioning marketplace.
WIN Principles for Generative AI are focused on establishing a joined-up approach globally and sit alongside other developments which share the same objectives and are supported by WIN members, see more here.
WIN Principles for Generative AI are available in English, Spanish, French, Korean and Portuguese.
