Donna Nagda Accuses AI Labs of Regulatory Capture
Donna Nagda claims major AI laboratories are slowing model releases to protect corporate interests against open-source competition rather than for safety reasons.
AI researcher and physician Donna Nagda argues that major AI laboratories are implementing pacing to slow the release of new models. She asserts this strategy is not driven by safety concerns but is an attempt at regulatory capture designed to protect corporate interests against the rapid growth of open-source AI.
Nagda suggests that large labs are increasingly threatened by the performance of smaller, sector-specific models and the success of open-source initiatives, particularly those emerging from China. She points to the record acquisition of Hugging Face by Nvidia as evidence of a strategic shift to counter these open-source threats.
While Nagda advocates for the acceleration of AI development to ensure the United States remains competitive, she does not oppose all restrictions. She supports the establishment of international standards and guardrails for high-risk applications, specifically citing defense and biology, with a focus on gain-of-function research.
