Deepfake images involving children, what realistic safeguards could actually help?

I’m adding this link from OpenAI:

https://openai.com/index/introducing-child-safety-blueprint/?utm_source=chatgpt.com

I highly recommend everyone that is interested to read the document in that link.