diff --git a/.changeset/giant-donkeys-wink.md b/.changeset/giant-donkeys-wink.md
new file mode 100644
index 0000000000..c8864127ee
--- /dev/null
+++ b/.changeset/giant-donkeys-wink.md
@@ -0,0 +1,5 @@
+---
+"@khanacademy/perseus": patch
+---
+
+[Image] | (a11y) | Add aria-describedby to Explore Image modal
diff --git a/packages/perseus/src/widgets/image/components/explore-image-modal-content.tsx b/packages/perseus/src/widgets/image/components/explore-image-modal-content.tsx
index 9a7004b542..7011f793a2 100644
--- a/packages/perseus/src/widgets/image/components/explore-image-modal-content.tsx
+++ b/packages/perseus/src/widgets/image/components/explore-image-modal-content.tsx
@@ -16,7 +16,12 @@ import type {CommonImageProps, ZoomProps, GifProps} from "./image-info-area";
 
 const MODAL_HEIGHT = 568;
 
-type Props = CommonImageProps & ZoomProps & GifProps;
+type Props = CommonImageProps &
+    ZoomProps &
+    GifProps & {
+        captionId: string;
+        longDescId: string;
+    };
 
 export default function ExploreImageModalContent({
     backgroundImage,
@@ -30,6 +35,8 @@ export default function ExploreImageModalContent({
     labels,
     range,
     zoomSize,
+    captionId,
+    longDescId,
 }: Props) {
     const [isGifPlaying, setIsGifPlaying] = React.useState(false);
     const context = React.useContext(PerseusI18nContext);
@@ -146,7 +153,10 @@ export default function ExploreImageModalContent({
                 )}
 
                 {caption && (
-                    <div className={styles.modalCaptionContainer}>
+                    <div
+                        id={captionId}
+                        className={styles.modalCaptionContainer}
+                    >
                         {/* Use Renderer so that the caption can support markdown and TeX. */}
                         <Renderer
                             content={caption}
@@ -165,12 +175,14 @@ export default function ExploreImageModalContent({
                     {context.strings.imageDescriptionLabel}
                 </Heading>
                 {/* Use Renderer so that the description can support markdown and TeX. */}
-                <Renderer
-                    content={longDescription}
-                    apiOptions={apiOptions}
-                    linterContext={linterContext}
-                    strings={context.strings}
-                />
+                <div id={longDescId}>
+                    <Renderer
+                        content={longDescription}
+                        apiOptions={apiOptions}
+                        linterContext={linterContext}
+                        strings={context.strings}
+                    />
+                </div>
             </div>
         </div>
     );
diff --git a/packages/perseus/src/widgets/image/components/explore-image-modal.test.tsx b/packages/perseus/src/widgets/image/components/explore-image-modal.test.tsx
index 78c4d06d98..0b1c2b4b74 100644
--- a/packages/perseus/src/widgets/image/components/explore-image-modal.test.tsx
+++ b/packages/perseus/src/widgets/image/components/explore-image-modal.test.tsx
@@ -234,6 +234,38 @@ describe("ExploreImageModal", () => {
         expect(screen.getByText("widget long description")).toBeInTheDocument();
     });
 
+    it("sets the describedby to short and long description IDs if there is a caption", () => {
+        // Arrange, Act
+        renderModal({
+            ...defaultProps,
+            backgroundImage: earthMoonImage,
+            caption: "widget caption",
+        });
+
+        // Assert
+        const dialog = screen.getByRole("dialog");
+        // Regex for "uniqueId-caption uniqueId-long-desc"
+        expect(dialog.getAttribute("aria-describedby")).toMatch(
+            /^:r\w+:-caption :r\w+:-long-desc$/,
+        );
+    });
+
+    it("sets the describedby to long description ID if there is no caption", () => {
+        // Arrange, Act
+        renderModal({
+            ...defaultProps,
+            backgroundImage: earthMoonImage,
+            longDescription: "widget long description",
+        });
+
+        // Assert
+        const dialog = screen.getByRole("dialog");
+        const ariaDescribedBy = dialog.getAttribute("aria-describedby");
+        // Regex for "uniqueId-long-desc"
+        expect(ariaDescribedBy).toMatch(/^:r\w+:-long-desc$/);
+        expect(ariaDescribedBy).not.toMatch(/caption/);
+    });
+
     describe("gif controls", () => {
         it("should render gif controls if the image is a gif", () => {
             // Arrange, Act
diff --git a/packages/perseus/src/widgets/image/components/explore-image-modal.tsx b/packages/perseus/src/widgets/image/components/explore-image-modal.tsx
index 2fc6948af0..5858e66a6b 100644
--- a/packages/perseus/src/widgets/image/components/explore-image-modal.tsx
+++ b/packages/perseus/src/widgets/image/components/explore-image-modal.tsx
@@ -14,6 +14,9 @@ type Props = CommonImageProps & ZoomProps & GifProps;
 
 export const ExploreImageModal = (props: Props) => {
     const context = React.useContext(PerseusI18nContext);
+    const uniqueId = React.useId();
+    const captionId = `${uniqueId}-caption`;
+    const longDescId = `${uniqueId}-long-desc`;
 
     const titleText = props.title || context.strings.imageAlternativeTitle;
     const title = (
@@ -39,7 +42,16 @@ export const ExploreImageModal = (props: Props) => {
         >
             <FlexibleDialog
                 title={title}
-                content={<ExploreImageModalContent {...props} />}
+                content={
+                    <ExploreImageModalContent
+                        {...props}
+                        captionId={captionId}
+                        longDescId={longDescId}
+                    />
+                }
+                aria-describedby={
+                    props.caption ? `${captionId} ${longDescId}` : longDescId
+                }
                 styles={{
                     root: wbStyles.root,
                 }}