Google Bard AI Can Now Retrieve Images, And That's Just The Start

Google has announced that a major new feature has arrived in its Bard AI tool, and though the platform still isn't as robust as Microsoft's Bing Chat — which leverages OpenAI's tech — it is quickly catching up with its competitor. Starting now, some Google users can ask Bard to retrieve images for them, and the AI will do so, eliminating the need to manually scroll through Google Images results.

There are some limitations at this time, including the fact that Bard is only delivering images in conversations written in English. However, Google says it will be adding support for other languages soon, and that ultimately this is only the start of Bard's image-related capabilities. 

As recently announced during the Google I/O 2023 keynote, Adobe's Firefly tech will be implemented directly into Bard, enabling it to generate (rather than simply retrieve) images. Microsoft already offers image-generating capabilities in Bing Chat using DALL-E 2.

The Bard image retrieval ability arrived on May 23, 2023, and it's able to handle fairly complex and highly specific tasks presented in plain language. At the most basic level, you can tell Bard to find you an image of something and it will, including a link back to the source of the image. That's not terribly useful, however, which is where more complex prompts come into play.

Users can ask Bard to, for example, retrieve five images of black cats with green eyes wearing gold collars, and the AI will scour through search results, plucking out five examples of images that meet your criteria. You can also ask it to find images based on their license — for example, to retrieve images that are in the public domain. It's necessary to double-check the license yourself, as Bard has been known to get its info wrong, but the capability is a welcome one regardless.

Beyond that, you may see images included with results even if you didn't explicitly ask for them. This could include, as an example provided by Google, images of food items included alongside details about restaurants, giving users that crucial visual element alongside the text. Google hasn't provided a specific release date for the Firefly-based generative image tool, but it's likely not far behind.