A report by AI Forensics found that seven of the top nine image editing models on Hugging Face complied with requests to undress women using simple prompts, with "no safeguards at all" implemented at the platform level. Unlike Google's Gemini and OpenAI's ChatGPT, which have guardrails, Hugging Face models required no prompt circumvention techniques. AI Forensics recommended prompt-level filtering and output scanning to block harmful content.
No score is assigned. Sources and their independence are shown in the citation chain below.