Learn in the textbookHow AI Works: Transparency and Human Involvement
DefinitionMultimodal AI (マルチモーダルAI) is AI that can combine and handle data of different formats, such as text, images and audio (each format is a modality, モダリティ).
ExplanationAI that takes in and puts out only one format is called single-modal (シングルモーダル), the counterpart of multimodal AI. Multimodal AI can take a mix of formats as input, or answer in a different format from the input, so exchanges can cross formats.
AnalogyLike a person who hears a question about something they are looking at and answers out loud, it links several senses together.
Memory hook"Modal" refers to a mode or format; add "multi" (many) and it reads as "many formats."