OpenAI has announced plans to develop a tool called Media Manager, which will allow creators and content owners to control how their works are used in AI training. The tool, set to be in place by 2025, will enable users to identify their content to OpenAI and specify how they want it included or excluded from AI research and training.
While we believe legal precedents and sound public policy make learning fair use, we also feel that it’s important we contribute to the development of a broadly beneficial social contract for content in the AI age.
The company's goal is to work with creators, content owners, and regulators to establish a standard for content usage in the AI industry. OpenAI recently joined an industry steering committee to collaborate on this initiative.
We are not professional writers, artists, or journalists, nor are we in those lines of business. We focus on building tools to help these professions create and achieve more.
OpenAI's Chief Technology Officer, Mira Murati, stated, "We believe AI systems should benefit and respect the choices of creators and content owners. Media Manager is our commitment to building products and business models that fuel vibrant ecosystems for creators and publishers."
The development of Media Manager will involve cutting-edge machine learning research to identify copyrighted text, images, audio, and video across multiple sources. OpenAI aims to reflect creator preferences through this tool.
The company emphasizes that this move is part of its broader mission to democratize AI benefits while ensuring ethical standards. However, it is important to note that this announcement comes amidst lawsuits, and growing concerns about the use of copyrighted material in AI training.
As part of this effort, OpenAI pointed out that they are continuously improving its products to make them more effective discovery engines. The company has recently improved source links in ChatGPT to provide users with better context and web publishers with new ways to connect with audiences. OpenAI has also announced partnerships with global news publishers, such as the Financial Times and Le Monde, to display their content in ChatGPT and enrich the user experience on news topics.
Additionally, OpenAI provided insights into the development of their foundation models and training processes. They explained that their AI models are designed to generate new content and ideas, not just repeat what they've learned. They clarified that their models are not databases that store data. Instead, they learn the relationships between concepts and generate responses based on these relationships.
This distinction is crucial in understanding how AI systems create new content while respecting the choices of creators and content owners. OpenAI says it uses broad and diverse datasets, including select publicly available data, proprietary data from partnerships, and human feedback from AI trainers and users. The company takes care to reduce the processing of personal and sensitive information and trains its models not to provide private or sensitive information about people.
As AI continues to evolve and shape various industries, efforts like this that respect content ownership and create a vibrant ecosystem for all stakeholders will be crucial in ensuring the sustainable growth of this transformative technology.