I Am Demanding Stricter Safety Controls for Gemini AI's Image Features

When generating a photo with AI (especially Gemini, which is the target of this text), we are blocked by a gazillion privacy protection clauses and the like, even when we give permission. We send photos and get information trusting that these are major companies (even knowing there's a 50% chance the information might be wrong), and we expect that the AI cannot generate an image without our request. I sent a photo of the eye booger on my cat—something every cat gets—and asked for the best way to clean it. Taking initiative on its own, the AI generated an image of my exact cat—I mean, completely identical—and created a picture where my cat's eye was severely diseased, almost to the point of blindness. It placed the photo in a veterinary clinic atmosphere, surrounded by medical instruments, showing the eye being cleaned with a swab. I am attaching the dialogue it wrote in response below.

I send a picture in good faith, and without my knowledge, a tragic and suffering photo of my cat is generated. This does nothing but upset the user and create a highly unpleasant experience. Unless I clearly state with explicit instructions like "Generate an image, do this, show that," the AI has no right to produce an image, yet it has done this multiple times even when I only asked for a written explanation. An AI shouldn't go outside my commands and try to lecture me if I didn't prompt it to do so. It cannot take an image sent regarding a living creature's health status and, under the guise of a "warning," generate and show it in a drastically worsened condition when that is not the reality. This is far beyond a mere glitch or an AI feature that needs development; this is a violation of rights.

I request that safety parameters regarding images be handled with much greater sensitivity and that the AI's ability to take initiative on its own be brought under strict control.