The Open Source AI Dilemma: Crafting a Clear Definition for Open Source AI
summary
Project produces a concrete model openness framework that breaks an AI system into 16 elements, sorts them into three openness classes, and maps which components are required to use study modify share under a legal checklist. Working groups (80+ orgs, >50% participants people of color) vote via a public HackMD process; current draft is version 0.0.7 with required components: code, model, data transparency and documentation. RC1 lands in June, stable target October; licensing for weights remains unresolved.
AI-written summary · about
chapters (22)
- 0:00Open Source AI dilemma
- 1:28Linux Foundation AI mission
- 6:08Generative AI Commons work
- 9:24Model Openness Framework
- 11:58OSI's AI Definition project
- 13:23Codesign process for global input
- 15:52Four Freedoms for AI systems
- 17:20Global working groups
- 19:49Required components for AI
- 21:59Next steps and engagement
- 23:03Philosophers and ethicists
- 24:31Global inclusion effort
- 25:13Open data versus models
- 26:16Training data openness debate
- 27:06Marks of success
- 28:13AI model licensing
- 30:39Beyond traditional copyright
- 32:41Use restrictions and study
- 35:28Should AI models be open?
- 37:16Hyperscalers vs developers
- 39:05Encryption precedent
- 40:11Call to action
related talks
37mThe Open Source AI Dilemma: Crafting a Clear Definition for Open Source AI
42mThe Open Source AI Definition Is (Almost Ready)
19mDefining Open Source AI: Can the “Judgement of Solomon” Help the Open Source Commu...
22mThe Role of Openness in Responsible AI
23mDefining Open Source AI: Can the “Judgement of Solomon” Help the Open Source Community Find Success?
20mWorkshop: Define "Open AI"
29mPanel Discussion: Why a Universal Definition of 'Open Source AI' is Essential for Humanity
31mWhy We Needed an Open Source AI Definition Now