What you need to know
- A new leak suggests Google is planning to implement an AI image generator into its chatbot, Bard.
- Diving into the code surfaced an unrelease patch note, which stated the generator would start with “English prompts” and offer a “Generate More” button.
- Speculation speaks to the generator’s possible usage of Imagen 2, which debuted for Vertex AI back in December 2023.
A leak shows there’s more to come for Google’s AI chatbot, Bard, that might leverage a piece of software used by Vertex AI.
As suggested by Dylan Roussel on X, Google plans to work an AI image generator into Bard (via 9to5Google). The tipster discovered a small set of unreleased patch notes, which state “You can create captivating images for work, play, and anything in between with Bard.” This text-to-image generator would begin with “English prompts.”
Moreover, the unreleased notes state a “Generate More” button will let users download their images and discover more unspecified options.
Roussel adds this generator will tap Google’s text-to-image diffusion technology, Imagen 2. What’s odd is that the notes tucked in Bard’s code suggested the feature would arrive on January 18, however, that’s not the case.
Google updated Bard today (Jan. 18) without this rumored image generator. According to the company, it has expanded “Bard Extensions” to more languages, including Japanese and Korean, alongside English. Bard first gained extensions in September 2023, bringing users real-time information from Maps, YouTube, Hotels, and Flights.
The second part of the latest update involves support for language export to Replit. Joining Python, the “Export to Replit” feature will support coders looking to transfer programs from C++, Javascript, Ruby, SQL, and Swift.
Back to the future? 👀Here’s what’s coming next in Bard. . . tomorrow.Image generation with Bard will use Imagen, Google’s Text-to-Image “diffusion technology.” The content of this changelog may still be changed until officially released. pic.twitter.com/9yRNjHNrLqJanuary 17, 2024
In October 2023, Google Search picked up its own AI image generator through the SGE Labs experience in October 2023. Starting with “draw” in the search bar, users could deliver a prompt to the software and receive four drawings in return.
Soon after, in December, Google debuted Imagen 2, a text-to-image generator for Vertex AI. The software can create “photorealistic” products alongside its text rendering support. The latter gave users more control over text prompts to get specific with what they wanted to create.
If Bard is indeed preparing to piggyback off Imagen 2, this could aid another recently leaked feature in the works. A lengthy leak suggests Bard will soon let users create bots and gain a “Tasks” function. This feature is said to “manage long-running tasks, such as generating images.”
More beyond that is unknown at the moment, as most of these features appeared quite barebones when they surfaced in the chatbot’s code.