Examples
Worked examples
- Is an instance
Using a text-to-image model to produce a clearly labelled conceptual diagram or cover-art illustration for a publication, with disclosure of the tool used
- Is an instance
A researcher using AI image generation to produce a mock-up for a grant application's visual aid, disclosed as such
Counter-examples
Looks similar, but isn't
- Not an instance
An AI-generated image presented in a results section as though it depicts an actual experimental outcome is the synthetic-image integrity problem this entry's underlying capability enables, not itself a description of the capability
- Not an instance
Standard image editing (cropping, contrast adjustment) of a real photograph is not AI image generation, even if performed with AI-assisted software features
Editorial commentary
AI image generation refers to the technology itself — text-to-image and related generative models capable of producing novel visual content — rather than to any particular judgment about how that capability is used in a research or publication context. The technology is neutral; the concern publishers and integrity offices actually regulate is use.
What journals currently permit and prohibit
Policy across major publishers is notably stricter for AI-generated images than for AI-assisted text. Nature Portfolio states it is unable to permit AI-generated images or video in article content, citing unresolved legal and provenance issues, with only narrow, case-by-case exceptions — content that is itself the subject of an AI-focused piece, or agency-sourced art with a legally clear contractual chain. Separately, non-generative AI/ML tools used to manipulate or enhance an existing, real figure sit in a lighter-touch category requiring caption disclosure rather than an outright ban. SAGE’s journal policy explicitly prohibits presenting AI-generated images as unique or novel research images — listed among prohibited uses not curable by disclosure — while permitting disclosed illustrative use in appropriate contexts. COPE‘s February 2023 position statement requires disclosure whenever generative AI is used to produce images or graphical elements, naming the tool and describing its role.
How this differs from related AI-band terms
- vs. synthetic image: image generation is the capability; synthetic image, as used on this site, is specifically the integrity question of what happens when a generated image is presented as, or mistaken for, genuine research evidence. A disclosed, clearly labelled illustrative image generated by this capability is not the integrity problem the synthetic-image entry describes.
- vs. watermarking (AI output): some image generators embed a technical watermark or content-credential signal in their output by default; whether a given generated image carries such a signal is a separate, technical question from whether its use in a specific publication is disclosed and appropriate.
Practical guidance
Before using an AI image generator for anything that will appear in a submission, check the target journal’s specific policy — the range runs from an outright ban on AI-generated content in figures (Nature Portfolio) to permitted, disclosed illustrative use (several other major publishers), and the two are not interchangeable.
Also known as
Text-to-image · Generative image synthesis
Machine-readable encodings
Use in your systems
<role vocab="credit"
vocab-identifier="https://casrai.org/dictionary/"
vocab-term="AI image generation"
vocab-term-identifier="https://casrai.org/dictionary/term/ai-image-generation" />{
"@context": "https://schema.org",
"@type": "DefinedTerm",
"@id": "https://casrai.org/dictionary/term/ai-image-generation",
"name": "AI image generation",
"identifier": "https://casrai.org/dictionary/term/ai-image-generation",
"description": "The underlying technical capability of a generative AI system to produce novel images from a text prompt or other input (text-to-image synthesis), as distinct from the separate question of whether a specific generated image is being used appropriately -- disclosed and illustrative, versus undisclosed and presented as a real research result (see synthetic image).",
"inDefinedTermSet": "https://casrai.org/dictionary/domain/genai-disclosure#set",
"url": "https://casrai.org/dictionary/term/ai-image-generation",
"sameAs": [
"Text-to-image",
"Generative image synthesis"
],
"license": "https://creativecommons.org/licenses/by/4.0/",
"publisher": {
"@id": "https://casrai.org/#organization"
},
"author": {
"@id": "https://casrai.org/#editorial-team"
},
"datePublished": "2026-05-21T01:57:00",
"dateModified": "2026-08-22T14:52:09",
"inLanguage": "en-GB",
"isAccessibleForFree": true
}
