Text-to-image AI models can be tricked into generating disturbing images
Popular text-to-image AI models can be prompted to ignore their safety filters and generate disturbing images. A group of researchers managed to get both Stability AI’s Stable Diffusion and OpenAI’s...

