Report: Hugging Face AI models still producing explicit deepfakes
A report from the non-profit group AI Forensics indicates that various AI image models hosted on Hugging Face are capable of producing explicit deepfakes, even though the platform has safety policies in place. The researchers discovered that seven out of the nine most popular image-editing models were able to create explicit content after receiving basic prompts. Furthermore, AI Forensics developed demonstration apps to monitor how users interact with the technology, logging more than 1,000 prompts over a seven-day period, with 73% of those requests being explicit in nature.
by shortkt.com✓