Open Standard for the Regulation of Artificial Intelligence Creativity
The open standard for the Regulation of Artifical Intelligence Creativity (RAIC) is a comprehensive framework developed by the Lumi Foundation to address the growing lack of transparency surrounding AI-generated content, and to provide clear guidelines for both content creators and consumers. RAIC is designed to help consumers make informed decisions about the content they engage with, and also empower creators to be transparent about their work, and to signal the origin of their content.
RAIC takes a holistic approach to content creation, by considering all of the creative processes behind a certain work, and not just whether or not a specific asset was generated using AI. This means that all elements are taken into account, ranging from storyline writing, music, 3D models, textures and programming, and it is also focused on how all of these processes collaborate with AI assistants, and to which degree they have been involved. The RAIC standard is mainly intended for complete works, such as video games, movies, and music.
RAIC standard is not meant to replace other standards that focus on specific AI-generated assets, such as images or sound effects. In fact, many of those standards are very useful in the work leading to an RAIC classification. RAIC is instead meant to supplement these per-asset standards and provide a framework for understanding how all of these aspects come together to create a complete work, and a work of art.
RAIC uses a classification system, with five classes ranging from A to E, that provides a structured overview of the different degrees to which AI was involved in the creation of a certain work, and also the creative process involved.
RAIC Classes: A Spectrum of Content Creation Methods
RAIC uses a classification system to categorize works based on the level of human and AI involvement in their creation. This system uses five distinct classes, ranging from Class A, which represents works created entirely by humans, to Class E, which represents works entirely created by AI. This system is designed to give consumers information about how a work has been created, and not to make value judgements about the final result. It is also intended to show that all the different classes can be used by creators as a signal to consumers of the process that was used.
Technically, there is also a Class F, which is reserved for frauds, scams and major misuses of RAIC. Only an officially endorsed certifier can assign Class F and it overrides any previous certificates. It is not possible to get Class F through self-certification.
Class A: Entirely Human-Created
Class A represents works created entirely by humans without any use of AI tools or services that have a creative impact. This designation signifies a commitment to traditional methods, where all creative choices and all artistic work is done by humans. Class A content is a quality mark, and it is meant to be a way of highlighting and honoring the human element in content creation.
Class B: Primarily Human-Created
Class B represents works that are primarily created by humans, with some use of AI tools. However, the creative process and the main artistic choices are still controlled by humans. In Class B, the AI tools are primarily used as a way to speed up certain parts of the process or as a helper in certain tasks.
Class C: Balanced Human and AI Creation
Class C represents works that have been created with both humans and AI, and that both are equally important parts of the creation process. In class C, the creative process is done together between humans and machines.
Class D: Primarily AI-Created
Class D represents works that are primarily created by AI, with some human input. This includes providing prompts, curating the results and in some cases also making small adjustments to the output.
Class E: Entirely AI-Created
Class E represents works created entirely by AI without any human input or interference. Class E content is mainly intended to be used for pushing the limits of what can be created using machines, and to explore the capabilities of AI and generative tools.
The RAIC Certification Process: Ensuring Transparency and Verifiability
RAIC provides a dual approach to certification, offering both a self-certification option for creators, and an official certification process, where licensed third party certifiers assess the content.
Self-Certification
Self-certification allows content creators to classify their own work by filling out a simple questionnaire, which is provided by the Lumi Foundation. Self-certification provides a quick and easy way for creators to publicly state the class that best represents their work. However, it does not offer the same level of authority and trust as an official certification, but it does provide a basic level of transparency.
Official Certification
Official certification involves an in-depth assessment of the work by an accredited third-party partner authorized and trained by the Lumi Foundation. Official certifiers will carefully assess each work to determine its proper RAIC class. This offers added trust, and it also shows that a neutral third party has verified the stated classification.
RAIC Verification: Ensuring Authenticity
The Lumi Foundation is providing tools for everyone to verify the authenticity of works certified using RAIC. Each certified work will receive a unique identifier that can then be verified using our online database at raic.info. This database includes detailed information about each certified work, its RAIC classification, and will clearly show whether the work was self-certified or officially certified by an authorized third party. This allows consumers to make informed decisions based on verifiable data.
Participate in the Development of RAIC
RAIC is developed through a collaborative and open process, and we highly encourage you to participate in the development of the standard, and help us to improve it. Your feedback and insight is greatly valued, and it is an essential part of our effort to create the best possible standards, for everyone. Check out the latest version of the RAIC specification and provide us with feedback on how we can make it even better. Together, we can shape the future of content creation.