The MetaverseUA Chair Newsletter #3 – November 2023

For sure, the creation of this tool is not a cause but rather a consequence of the predominantly negative stance taken by most artists regarding the use of their works by AI systems. You can see this general attitude in the bunch of legal battles popping up against AI models lately. A notable example is the case brought against Stability AI, Midjourney, and DeviantArt before the Californian courts. Here, the judge dismissed several claims, including those alleging copyright infringement by AI outputs and unfair competition, but permitted the plaintiffs to revise their primary complaint about the use of their images in Stability AI’s training via an amended lawsuit. We will thus have to wait for the outcome of the case, which in any case perfectly exemplifies the existing tensions and the difficulty of striking a balance between the interests of all parties involved.

Also interesting in terms of artists’ position on artificial intelligence are the guidelines published by the MPA (an association that represents the top five U.S. film studios, along with the video streaming service Netflix), which offer a somewhat more flexible and welcoming stance. In their statement, the MPA contends that existing doctrines and principles, like «fair use,» already provide courts and the US Copyright Office with adequate tools to handle emerging AI-related challenges. Additionally, they suggest shifting from a strict, unbending rule that wholly discourages AI usage in the industry, acknowledging the potential beneficial outcomes that it can offer (for example, by training an AI tool with a given artist’s copyrighted material to detect potential infringements). A summary of the guidelines can be found here.

AI news does not stop here, though. In this regard, it is crucial to highlight that the G7 leaders have recently agreed on a set of international guiding principles on AI and a voluntary Code of Conduct for the developers of such technology. 

The so-called Hiroshima AI Process includes four key elements:

1. An assessment of the primary risks, challenges, and opportunities associated with generative AI.
2. The Hiroshima Process International Guiding Principles, applicable to all entities in the AI ecosystem.
3. The Hiroshima Process International Code of Conduct for organizations involved in developing advanced AI systems.
4. A proposal for collaborative projects to support the creation of responsible AI tools and establish best practices.

If you want to read the text of the proposal in more detail, it can be found here. 

Also regarding governmental initiatives for responsible and regulated AI development, Spain’s notable step involves the establishment of a controlled testing environment (commonly referred to as a «sandbox») aligned with the European Artificial Intelligence Regulation. Indeed, through a Royal Decree enacted on November 8, a call has been issued to Spanish companies interested in participating in this initiative. The primary goal of this project is to enhance communication between users and AI system suppliers, as well as between these suppliers and relevant authorities. The aim is to facilitate the exchange of information, promote a deeper understanding of compliance requirements, ensure transparency, encourage the development of reliable systems, and conduct testing within a controlled environment, so as to gather information and elaborate technical guides that can provide assistance in this domain.