Safeguarding generative AI. (Part 1)