In the next period of time, Google is expected to add a large number of artificial intelligence features to its various products, some of which are rumored to be announced at the I / O 2023 developer conference.
According to the latest information discovered by foreign media 9To5Google, Google seems to be building a text-to-image generator in the Android version of the Gboard mobile input method.
Foreign media decompiled the Gboard Beta App (12.7.05.507749191) uploaded by Google to the Play Store, and found that some strings in the code referenced “Imagen Keyboard”.
Imagen is Google’s text image model. It claims that just by giving a description of a scene, Imagen can generate high-quality, high-resolution images, a bit like OpenAI’s DALL·E.
According to the code content, the “Imagen Keyboard” function will appear in the shortcut page. Therefore, it is conceivable that users can input text in the “Imagen Keyboard” and directly generate pictures for writing, chatting, etc. It’s unclear when the feature will roll out.