Hi OpenAI team and community,
I’m a writer and visual storyteller who uses ChatGPT to bring fictional narratives to life—especially emotionally rich moments around family, birth, and parenthood. I’ve been experimenting with image generation to support those stories, and I’ve noticed something confusing and a bit frustrating.
Some prompts that involve respectful, non-graphic scenes of childbirth or newborn care (e.g., a father holding a newborn, a mother in labor, skin-to-skin bonding) are blocked due to content policy. Others with nearly identical tone and detail go through just fine. For example, I successfully generated an image of a father holding his newborn in a hospital room, but a very similar scene of him catching the baby at birth was flagged.
I completely support OpenAI’s safety guidelines, but I’d love to see more nuance and flexibility when it comes to emotionally grounded, artistic representations of birth and family connection—especially in clearly fictional and respectful contexts. These moments are meaningful, human, and powerful,