Following the news that Spotify has also joined the AI playing field by launching an AI DJ that will customize your playlist according to your listening patterns as well as commentate on the songs giving you some little titbits about them, it is increasingly becoming clear that 2023 will be the year of AI.
According to 9to5Google, the next AI product will likely come from Google, going by the source code the outlet has come across in a recent teardown of the Gboard APK. In the code, there are lines that refer to an “Imagen Keyboard”. Imagen on its own is an AI system that is able to create images from input text.
This text-to-image feature on Gboard will reportedly appear in the shortcuts strip, where the Clipboard and One-handed mode is normally found. Keep in mind though that despite appearing in the source code, it is not a guarantee that the tech giant will roll out the feature in the next update.
It goes without saying that there is no concrete date as to when Google will launch this text-to-image generator. Once they do, it will go head-to-head against DALL-E2 which is also a text-to-image generator created by Open AI, also the creators of the popular ChatGPT.
According to an interesting study carried out by Google, human raters preferred the images developed by Imagen, over those developed by other platforms including VQ-GAN+CLIP, Latent Diffusion Models, and DALL-E 2 in side-by-side comparisons in terms of sample quality and image text alignment.
In addition to Bard, Google’s answer to ChatGPT and now Imagen Keyboard, the AI party at Google is just getting started. There are reportedly over 20 AI products that the tech giant is working on, including a wallpaper maker and a tool that summarizes videos by coming up with a new one.
