Review Multimodal

Multimodal describes a model or system that can work with more than one kind of data, such as text, images, audio, video, or sensor information.

A multimodal system may accept one type of input and produce another, or combine several types at once. The exact supported combinations and quality differ by product and model.

Combining modalities can make tools more useful, but it also creates new privacy, provenance, accessibility, and evaluation questions across each input and output type.

A user supplies a photograph and a spoken question, and the system returns a text explanation of what appears in the image.

Multimodal does not mean the system handles every medium equally well or combines them with humanlike perception.

More than one kind of input or output is in play.

* [[generative-ai|Generative AI]] * [[large-language-model|Large Language Model]] * [[data-provenance|Data Provenance]] * [[embodiment-premium|Embodiment Premium]]

* [NIST, *Adversarial Machine Learning: A Taxonomy and Terminology of Attacks and Mitigations* (NIST AI 100-2 E2025)](https://doi.org/10.6028/NIST.AI.100-2e2025) — defines multimodal models as processing and relating information from multiple sensory modalities.

* Review media lineage with [[data-provenance|Data Provenance]] and generated content with [[generative-ai|Generative AI]].