Linux Foundation’s latest initiative aims to promote ‘irrevocable’ open-source AI models

The Linux Foundation today welcomed a new initiative that aims to encourage the development of high-quality, openly licensed artificial intelligence models that will be free to use and “push creativity forward,” in response to a growing demand for truly open AI.

The Open Model Initiative or OMI was established in response to a number of recent decisions by creators of popular open-source models to alter their licensing terms. The creators highlighted the recent licensing change announced by Stability AI Ltd., regarding its popular image-generation model Stable Diffusion 3 (SD3). That model had previously been entirely free and open, but the changes introduced a monthly fee structure and imposed limitations on its usage.

Stability AI was also criticized for the lack of clarity around its licensing terms, but it isn’t the only company to have introduced licensing restrictions on previously free software.

The OMI intends to eliminate all barriers to enterprise adoption by focusing on training and developing AI models with “irrevocable open licenses without deletion clauses or recurring costs for access,” the Linux Foundation said.

The initiative was formed by a number of organizations, including Invoke AI Inc., CivitAI Inc. and Comfy Org, and has the support of community projects such as the Sentient Foundation and Wand Synthesis AI Inc. It has been granted the status of an “official open-source foundation” hosted by the Linux Foundation, and governance will be led by a community-based Steering Committee.

According to the OMI, its primary objective is to facilitate the development of truly “open-source, capable, and ethical” generative AI models. It has outlined a plan to achieve this, saying it will first establish a governance framework and working groups to foster collaborative community development, and conduct a survey that aims to gather feedback on future model research and training from the AI community.

The OMI’s plan also calls for the creation of “shared standards” that will enhance AI model interoperability and metadata practices, and the development of a transparent, open-source dataset for AI training that any developer can use. In addition, it’s looking to create an alpha test model for targeted red-teaming, and release the first version of that model, along with fine-tuning scripts, before the end of the year.

Linux Foundation Executive Director Jim Zemlin said the organization is committed to fostering open and collaborative development around AI. “With the Open Model Initiative, we are taking a significant step toward making AI accessible and beneficial for everyone, building an environment where creativity and progress in AI can thrive without barriers,” he promised.

Kent Kiersey, founder and chief executive of Invoke, stated his belief that open-source remains the best path forward for AI that will benefit everyone. “Through OMI’s commitment to open source development under the Linux Foundation, we believe the collective will deliver high-quality, competitive models with open licenses that push AI creativity forward, are free to use, and meet the needs of the community,” he said.

The OMI is open for anyone who is passionate about open-source software and AI, and has extended an invitation to get involved via its official GitHub repository and Discord channel.

Source