Model: openrouter/google/gemini-2.5-flash-preview:thinking
-- none --
https://eur-lex.europa.eu/legal-content/EN/TXT/HTML/?uri=OJ:L_202401689
Summary of mentions of open source here, including what the document defines open source to mean
Based on the provided text of Regulation (EU) 2024/1689 (the "Artificial Intelligence Act"), here is a summary of mentions of "open source", including how the document defines or describes it:
-
General Mention and Encouragement:
- Recital (89) states that third parties making accessible "tools, services, processes, or AI components other than general-purpose AI models" under a free and open-source licence should not be mandated to comply with upstream provider responsibilities towards those who integrate them. It also encourages developers of such resources to implement documentation practices like model cards and data sheets to promote trustworthy AI.
- Recital (102) acknowledges that software and data, including models, released under a free and open-source licence can contribute to research and innovation and provide growth opportunities.
-
Document's Description/Definition of Free and Open-Source Licenses:
- Recital (102) describes a free and open-source licence as one that:
- Allows software and data (including models) to be openly shared.
- Where users can freely access, use, modify and redistribute them or modified versions thereof.
- A license can still be considered free and open-source even if it includes conditions such as the original provider being credited or identical or comparable terms of distribution being respected.
- Recital (103) clarifies that "Free and open-source AI components" cover the software and data (including models), tools, services or processes of an AI system. It specifies that such components are not covered by the open-source exceptions in the Regulation if they are provided against a price or otherwise monetised (e.g., via technical support, services through a platform, using personal data for reasons beyond security/compatibility/interoperability), unless the transaction is between microenterprises. Making components available through open repositories does not, by itself, constitute monetisation.
- Recital (102) describes a free and open-source licence as one that:
-
Key Provisions Regarding Open Source AI Systems/Models:
- General Exemption from Scope: Article 2(12) is a key provision stating that the Regulation does not apply to AI systems released under free and open-source licences, unless they are placed on the market or put into service as:
- High-risk AI systems (under Article 6).
- AI systems that fall under the transparency requirements of Article 50 (e.g., systems intended to interact with humans, emotion recognition, biometric categorisation, deepfakes).
- Limited Exemption for AI Components in the Value Chain: Article 25(4) states that third parties supplying AI tools, services, components, or processes, other than general-purpose AI models, under a free and open-source licence are exempted from certain obligations applicable to parties in the AI value chain (specifically, the requirement of a written agreement specifying necessary information and access).
- Exemptions for Open-Source General-Purpose AI Models (with conditions):
- Article 53(2) provides an exception from the obligations for providers of general-purpose AI models regarding technical documentation (Art 53(1)(a)) and providing information to downstream providers (Art 53(1)(b)) if the models are released under a free and open-source license and their parameters (including weights), information on model architecture, and information on model usage are made publicly available. This exemption does not apply to general-purpose AI models with systemic risks.
- Article 53(4) states that providers relying on open-source models can use codes of practice to demonstrate compliance with their obligations, providing a presumption of conformity when harmonised standards are published.
- Article 54(6) mirrors the exemption in Article 53(2) regarding the obligation to appoint an authorised representative for providers of general-purpose AI models. This also does not apply to general-purpose AI models with systemic risks.
- No Exemption for Core Obligations: Recital (104) and Article 53 implicitly reinforce that even open-source general-purpose AI models are not exempt from fundamental obligations like putting in place a policy to comply with Union copyright law or drawing up and making publicly available a summary of the content used for training.
- General Exemption from Scope: Article 2(12) is a key provision stating that the Regulation does not apply to AI systems released under free and open-source licences, unless they are placed on the market or put into service as:
In summary, the regulation defines "free and open-source" based on the freedom for users to access, use, modify, and redistribute the software/data/models/components under specific licenses, with conditions like attribution being permissible. While generally exempting open-source AI systems and components from many requirements, it imposes specific obligations or removes exemptions for open-source systems/models classified as high-risk or general-purpose (particularly those with systemic risks), or those falling under specific transparency rules.
264,864 input, 3,337 output, {"completion_tokens_details": {"reasoning_tokens": 2288}, "cost": 0.0514091}
so it says specific licenses but does it clarify which specific licenses count?
No, the document does not list specific names of licenses that qualify as "free and open-source".
Instead of listing names (like GPL, MIT, Apache, etc.), the Regulation defines what kind of terms a license must grant for it to be considered "free and open-source" for the purposes of the Regulation.
According to Recital (102), such a license must essentially allow software and data, including models, to be:
- Openly shared.
- Where users can freely access, use, modify and redistribute them or modified versions thereof.
The Recital further clarifies that a license can still be considered free and open-source even if it includes standard conditions like requiring attribution (crediting the original provider) or respecting identical or comparable terms of distribution.
Additionally, Recital (103) adds a crucial point: even if a license meets these freedom criteria, the AI components released under it may not benefit from the open-source exceptions in the Regulation if they are provided against a price or otherwise monetised (e.g., through included technical support, services via a platform, or using personal data for reasons other than improving security/compatibility/interoperability), unless the transaction is between microenterprises.
So, the focus is on the freedoms granted by the license terms and the nature of the economic activity surrounding the provision, rather than a pre-approved list of license names.
265,927 input, 1,066 output, {"completion_tokens_details": {"reasoning_tokens": 756}, "cost": 0.04362005}