Paper Number
ICIS2025-1318
Paper Type
Complete
Abstract
The rapid increase of Generative AI (Gen AI) models advertised as “open source” has spurred criticism, as many of these models release only limited technical components and impose usage restrictions. This paper addresses this lack of clarity by developing a taxonomy and identifying archetypes of open-source Gen AI models. Drawing on open-source software (OSS) literature and emerging open-source Gen AI research, we construct a taxonomy comprising three domains, technical openness, governance, and licensing, spanning eight dimensions and eighteen characteristics. Applying this taxonomy to 248 real-world open-source Gen AI models, we perform cluster analysis to derive six archetypes, reflecting distinct configurations of openness. Our study contributes to OSS and open-source Gen AI research by introducing open-source Gen AI models as a novel class of open systems and offering a nuanced view of openness in this new context. Our findings provide actionable guidance for organizations navigating the open-source Gen AI landscape.
Recommended Citation
Diaferia, Lorenzo; De Rossi, Leonardo Maria; and GAUR, Aakanksha, "Making Sense of Open-Source Generative AI Models: A Taxonomy and Archetypes" (2025). ICIS 2025 Proceedings. 8.
https://aisel.aisnet.org/icis2025/gen_ai/gen_ai/8
Making Sense of Open-Source Generative AI Models: A Taxonomy and Archetypes
The rapid increase of Generative AI (Gen AI) models advertised as “open source” has spurred criticism, as many of these models release only limited technical components and impose usage restrictions. This paper addresses this lack of clarity by developing a taxonomy and identifying archetypes of open-source Gen AI models. Drawing on open-source software (OSS) literature and emerging open-source Gen AI research, we construct a taxonomy comprising three domains, technical openness, governance, and licensing, spanning eight dimensions and eighteen characteristics. Applying this taxonomy to 248 real-world open-source Gen AI models, we perform cluster analysis to derive six archetypes, reflecting distinct configurations of openness. Our study contributes to OSS and open-source Gen AI research by introducing open-source Gen AI models as a novel class of open systems and offering a nuanced view of openness in this new context. Our findings provide actionable guidance for organizations navigating the open-source Gen AI landscape.
When commenting on articles, please be friendly, welcoming, respectful and abide by the AIS eLibrary Discussion Thread Code of Conduct posted here.
Comments
12-GenAI