Open Shouldn't Mean Exempt: Open-Source Exceptionalism and Generative AI

Open-source AI should face the same accountability as proprietary systems. The paper argues that simply being "open" should not exempt generative AI from ethical or legal standards, challenging the assumption that open-source status automatically makes AI systems more responsible or trustworthy.

"Democratization" claims about open-source AI are often overstated. The author refutes the core argument that open-source GenAI disrupts corporate monopolies and democratizes access, finding instead that these systems can reinforce existing power structures rather than challenge them.

Open-source AI can enable harmful activities. The paper documents that open-source generative AI systems can facilitate unlawful conduct and exacerbate environmental harms, demonstrating that openness alone does not prevent misuse.

A narrow research exemption is preferable to blanket exemptions. Rather than exempting all open-source AI from regulation, the author proposes a carefully limited safe harbor specifically for legitimate, non-commercial research that meets strict compliance criteria.

"Innovation" and "openness" rhetoric shouldn't substitute for rigorous policy analysis. The paper warns that appeals to innovation and democratization are frequently used to justify regulatory exemptions without solid evidence, and that policy should be based on clear-eyed assessment of actual societal impacts rather than ideological claims.