Google has announced a significant update to Bard which lets you create free high-quality AI-generated images. Meanwhile, Google’s Gemini Pro model, announced for Bard last month, can now support over 40 languages and works in over 230 countries and territories worldwide.
According to Google, its new image generation tool in Bard is powered by their updated Imagen 2 model which balances quality and speed, while delivering high-quality and photorealistic outputs. This is the same model used by Galaxy AI for its AI-assisted photo editing features on the Samsung Galaxy S24 series.
How to create images on Google Bard?
Here’s how to generate your AI-generated images on Google Bard:
- Launch Bard on your browser (Desktop and Mobile)
- Type in your desired photo description in the prompt (e.g. Create an image of a cat surfing or generate an image of a delicious nasi lemak).
- By default, it will show two images, tap on “Generate More” to view more options
- You can tap on the download icon on the top right icon to get a higher-resolution image
Take note that the full-size image has a resolution of 1536×1535. You can also download all of the images by tapping on the share icon and selecting “Download all images”. The image generation feature is available in Malaysia but it only supports English for now.
You can refine your AI-generated image results by providing more details to Bard. For example, you can mention the desired theme or style (e.g. futuristic, abstract), specific objects or elements you wish to include, the mood or feeling and any colour preference (e.g. red car, blue house, orange cat).
You can generate text and images in a single prompt
That’s not all, you can ask Bard to generate content with images. If you want Bard to create a social media post with an image, you can submit prompts such as “Write social media post and generate a mouthwatering image of Nasi Lemak Ayam Goreng”. The generated results for social media post even included suggested hashtags for the topic.
Similarly, you can also ask Bard to generate long-form text content with generated images as shown above.
Google has added safety measures
To ensure safe and responsible use of Bard, Google is using SynthID to add digital watermarks into the pixels of the generated images. Bard will also not generate images that depict violence or harm to people or property, and it will suggest positive alternatives instead. There are also limitations on generating images of specific people without their consent which helps to prevent misuse of identity.
[ SOURCE ]
