OpenAI announces the development of a tool enabling content creators to “opt out” of AI training

OpenAI is in the process of developing a tool called Media Manager to empower creators and content owners to exercise greater control over how their content is utilized in training generative AI models. This tool will enable creators to identify their works to OpenAI and specify preferences regarding inclusion or exclusion from AI research and training datasets.

The aim is to have Media Manager available by 2025, with OpenAI collaborating with creators, content owners, and regulators to establish industry standards, potentially through the industry steering committee it recently joined. Media Manager will leverage cutting-edge machine learning research to identify copyrighted text, images, audio, and video across various sources and accommodate creator preferences.

The development of Media Manager reflects OpenAI’s response to mounting criticism regarding its approach to AI development, particularly its reliance on scraping publicly available data from the internet. Notably, OpenAI faced legal challenges, such as a lawsuit from prominent U.S. newspapers accusing the company of IP infringement related to its use of generative AI.

To address concerns and mitigate legal risks, OpenAI has taken steps to engage with content creators. This includes allowing artists to opt out of having their work included in training datasets, enabling website owners to control content scraping via robots.txt standards, and establishing licensing agreements with content providers. However, some creators argue that OpenAI’s existing solutions are insufficient, citing challenges with opt-out workflows and compensation.

Beyond OpenAI, other entities are also working on developing tools for managing content provenance and opt-out options in the context of generative AI. These efforts aim to balance innovation in AI research with respect for intellectual property rights and creator preferences.