OpenAI Promises Creators New Tool to Control How Their Content Trains Generative AI Models

OpenAI has introduced a tool that will supposedly let content creators decide whether and how their work is incorporated into the company’s generative AI model training and related research. The upcoming Media Manager feature will allow people to tag their work and specify what role it can play in OpenAI’s model development.

AI Media Manager

Public data scraped from the internet is at the heart of most generative AI models built by OpenAI and developers. The generative AI developers cite the fair use doctrine to explain their training data sources, but the practice has nonetheless led to plenty of controversy and more than a few lawsuits over copyright infringement and related allegations. Media Manager looks like an indirect answer to those complaints by making it possible for the people and companies who make the content to customize how their intellectual property is used by OpenAI or opt out entirely.

“We need an efficient, scalable solution for content owners to express their preferences about the use of their content in AI systems,” OpenAI explained in a blog post. “This will require cutting-edge machine learning research to build a first-ever tool of its kind to help us identify copyrighted text, images, audio, and video across multiple sources and reflect creator preferences. We’re collaborating with creators, content owners, and regulators as we develop Media Manager. Our goal is to have the tool in place by 2025, and we hope it will set a standard across the AI industry.”

As part of the data sourcing and use shift, OpenAI is upgrading its models to pull from more diverse data sources. This includes not only publicly available data but also proprietary data obtained through partnerships with companies and organizations.OpenAI claims these partnerships provide users with improved content discovery and engagement.

Most recently, Stack Overflow and OpenAI agreed to a deal that sees OpenAI helping Stack Overflow with its Overflow AI set of generative AI features while Stack Overflow’s validated technical data will become part of ChatGPT’s resources and help train future OpenAI models. That followed on the heels of a deal with the Financial Times (FT) to license its content to improve OpenAI’s models while bringing customized AI features to FT readers. It’s similar to the deals OpenAI has made with Axel Springer, the Associated Press, Le Monde, and others.

Follow @voicebotaiFollow @erichschwartz

Stack Overflow Fully Embraces Generative AI With OpenAI Partnership

OpenAI and The Financial Times Trade Content License for Generative AI Features

The New York Times Sues OpenAI and Microsoft for Training Generative AI Models With Copyrighted Work