Gemini AI Mobile App Inverts Image Generation Prompt Instructions

I have been using the Gemini AI mobile application since December 5, 2025. I am a manga artist, and I provide the app with story concepts while also sharing my own drawings with it. While I specifically expect the app to generate illustrations, it generates images even when I do not ask it to "create a photo," yet conversely, when I explicitly state "create a photo," it fails to do so. In short, it responds in exact opposition to my prompts.

When I give the command "create a photo," I expect it to produce an image, and when I say "do not create a photo," I expect it to refrain from generating one. This issue disrupts my workflow and prevents me from using the app efficiently. I demand that the app be updated to correctly interpret given instructions and deliver appropriate outputs accordingly.