OMI is working to gather feedback on what an open-source LLM should look like and is expected to release an alpha version of its model by the end of the year.
The Linux Foundation has adopted the Open Model Initiative (OMI) to maintain open-source status for large language models (LLMs) and AI tools. This move comes amid growing concerns over AI models that shift from open to restrictive, proprietary licensing, limiting access and innovation in the field.
Challenges of Restrictive AI Licensing
The OMI was founded by CivitAI, ComfyOrg, and Invoke in response to increasing restrictions on AI licensing. Recent AI models, such as Stability AI’s Stable Diffusion 3, transitioned from being fully open to requiring a monthly fee and usage limitations. The OMI aims to counteract such trends by promoting openly licensed AI models for image, video, and audio generation. Its key goals include:
- Preventing restrictive non-commercial license agreements.
- Avoiding compromised capabilities.
- Fostering collaboration and innovation.
- Preventing the concentration of power among a few corporations.
- Eliminating recurring costs and deletion clauses.
See also: CIO Insights: The Strategic Importance of RAG and Open Source in GenAI
Next Steps for Open AI Development
As an organization under the Linux Foundation, OMI is working on creating a governance framework and establishing working groups to drive community-driven development. The initiative plans to gather feedback on what an open-source LLM should look like and is expected to release an alpha version of its model by the end of the year. Meanwhile, the Open Source Initiative (OSI) is developing an official open-source AI code license, aiming to release its first version by October.
OMI encourages individuals and organizations passionate about open-source AI to participate in its efforts through platforms like Discord and GitHub. The goal is to develop high-quality, competitive AI models that remain free to use and accessible to all.
Related articles:
