Hugging Face

Throughout 2023-2024 Hugging Face co-founder Clement Delangue and policy lead Yacine Jernite were among the most visible commercial AI executives publicly endorsing the EU AI Act's open-source exemptions and model-card transparency requirements, contrasting sharply with OpenAI, Anthropic and Mistral lobbying positions. Hugging Face was the only major foundation-model hosting platform to publish detailed model evaluation and bias-disclosure tooling as default for hosted models, and to maintain ethical-use policies that include takedown commitments for non-consensual intimate imagery models.

Throughout 2023-2024 Hugging Face removed multiple hosted projects and models specifically designed for non-consensual deepfake image and voice generation, including high-profile takedowns following BBC and 404 Media investigations. The platform updated its content policy to explicitly prohibit non-consensual intimate imagery models and tools whose primary use is impersonation without consent. Critics including the EFF noted the policy enforcement is reactive rather than preemptive but is more rigorous than other model-hosting platforms.

In December 2023 Stanford Internet Observatory researchers disclosed that the widely-used LAION-5B image-text training dataset hosted on Hugging Face contained at least 1,008 identified child sexual abuse material (CSAM) images, with researchers estimating likely thousands more. Hugging Face and LAION removed the dataset within days of disclosure. The episode highlighted gaps in pre-hosting safety screening for very large training datasets and prompted Hugging Face to expand its dataset-content policy enforcement.