Answer
It is not a verified account of the resulting pixels. The concise accessibility label remains AI-generated image. Internal visual checks do not certify that every instruction was followed or automatically produce another image.
To ask about the actual result, leave @ai on, choose Text, reply to the image and ask for a description or a specific visible detail. For example, ask “What text can you read in this image?” or “Describe what is visible.” Existing image reading and its backup remain available.
Review generated images before relying on their details, especially text, labels and visual facts. Asking for an analysis in Text is different from requesting a new generation or edit in Image.