Okay, this is wild. Turns out some researchers managed to trick a major AI chatbot into spitting out seriously graphic and disturbing images with incredibly simple prompts – like just asking it to "restore a photo" when none was even there! Then they tweaked it slightly and got it to generate all sorts of violent and sexual content. It really makes you wonder about the effectiveness of current AI safety measures. The company behind it says they've added safeguards now, but the fact this even happened with such basic instructions is a huge red flag. It highlights how tricky it is to truly moderate what these powerful tools can create. Are we really prepared for what happens when these systems become even more advanced without foolproof controls? #AI #Technology #Safety https://indianexpress.com/article/techn… 🔗