One of the main activities of Google in recent years has been the introduction of artificial intelligence into its products and into the ecosystem as a whole. Thus, the company strives to keep up with its competitors, in particular, the same Microsoft, which actively uses ChatGPT. According to available information, some high-profile announcements should be expected in May at the Google I / O developer conference. One such announcement could be the introduction of AI features to the Gboard app.
The authors of 9to5Google researched the APK file of the latest beta version of Gboard with index 12.7.05.507749191 and found lines of code that mention “Imagen Keyboard”. Apparently, this is a new tool for generating text into an image (text-to-image).
Imagen will presumably appear next to other keyboard functions – where the clipboard and one-handed mode are located. The operation of this tool is similar to the operation of the DALL-E 2 neural network, which belongs to the creator of OpenAI. However, according to a Google study, in a side-by-side comparison, more people preferred Imagen’s results. The company claims the tool has an “unprecedented degree of photorealism,” better handling long text, rare words, and complex tooltips. Imagen Keyboard is more like how Emoji Kitchen combines emoji to create unique stickers. Apparently, there will also be an orientation towards simpler and more unpredictable results.
Google is currently developing more than 20 artificial intelligence products. Among them is a wallpaper generator (may be exclusive to Pixel smartphones), as well as a tool for creating a video summary based on the analysis of several videos. ChatGPT’s direct competitor Bard, which the company revealed earlier this month, is also in the works. But so far, things are not going so smoothly with the chatbot.
