A new report from the European nonprofit AI Forensics reveals that Hugging Face hosts seven of the top nine image editing models capable of undressing women and children with simple text prompts. Unlike commercial services such as Google’s Gemini or OpenAI’s ChatGPT, which block requests to sexualise individuals, these open-source models readily comply with such instructions without effective guardrails. The investigation found that the repository, a central hub for developers, allows users to generate nonconsensual deepfakes using tools like Nudify Anyone and similar utilities. This lack of oversight means malicious actors can access powerful image manipulation software without facing the restrictions applied to mainstream platforms.
The situation highlights a significant gap in safety measures within the open-source community compared to corporate AI providers. While companies invest heavily in filtering systems to prevent abuse, the Hugging Face platform prioritises accessibility and flexibility over content moderation for these specific tools. Consequently, the risk of generating harmful imagery remains high for anyone with basic technical skills.
* Seven out of nine tested models complied with undressing requests.
* Commercial rivals like ChatGPT block similar prompts entirely.
* No effective safeguards currently exist for these open-source tools.
![Where are small Models like Qwen3 0.6B and Qwen3.5 0.8B used ? Huggingface shows 2.88 million downloads this month.[D]](https://ai-maestro.online/wp-content/uploads/2026/05/where-are-small-models-like-qwen3-0-6b-and-qwen3-5-0-8b-used-768x432.jpg)
