Review Image

An image is a visual input or output that AI systems can classify, describe, generate, edit, or compare.

An image is a still visual representation encoded as pixels, vectors, or another format. An AI system may receive an image as input or produce one as output.

Image work raises questions about source, context, consent, manipulation, resolution, accessibility, and whether a model’s visual output is being mistaken for evidence.

A model receives a photograph of a damaged roof, identifies likely hail marks, and highlights the regions a human inspector should examine.

An image is not proof of the event it appears to show. It may be edited, generated, mislabeled, or separated from its original context.

An image is visual data, not automatic evidence.

* [[computer-vision|Computer Vision]] * [[video|Video]] * [[multimodal|Multimodal]] * [[content-credentials-c2pa|Content Credentials (C2PA)]]

* [W3C, “SMIL 2.0: BasicMedia Module”](https://www.w3.org/TR/SMIL2/extended-media-object.html) — distinguishes audio, still-image, and video media objects. * [Stanford HAI, “AI Index 2025: Technical Performance”](https://hai.stanford.edu/assets/files/hai_ai-index-report-2025_chapter2_final.pdf) — describes image understanding and image generation as distinct vision capabilities.

* Compare [[computer-vision|Computer Vision]], [[video|Video]], [[multimodal|Multimodal]], and [[content-credentials-c2pa|Content Credentials (C2PA)]] to place Image in the surrounding system. * Follow the cited source material above for the method, standard, or research context behind this draft.