SEE TOO Most Read
ON THE HIGH Tags
In the image there is a person holding a smartphone, on the screen there is the google symbol and the name "bard".
4 minutes reading

How to create images in Bard, Google's AI

wanessa alves avatar
Vanessa Alves
August 22th, 2024
The new update to Bard, Google's generative AI, allows users to create images. Check step by step.

"Hi, I'm Bard”. It is with a greeting that Bard welcomes users who seek out the network to use AI in text production and research. The site, considered a rival to the famous ChatGPT, now offers free access to creating images through AI (artificial intelligence). 

According to Google Support instructions, image creation requests must be made via text in the image itself. Bard. It is important to highlight that the image description must be written in English, because the site does not yet recognize commands in other languages. 

Image of a female astronaut generated in bard, google's generative AI.
Image generated in Bard

The creation of images through AI is something that has worried experts and the general population, because malicious users are using the technology to generate false images and publish them online. According to Google, Bard has a watermark feature incorporated into the pixels of the creations and this helps to discover whether the generated image is Bard's invention or not. 

Furthermore, the multinational states that the system has security techniques to prevent the creation of images of people linked to violent, offensive, nudity or sexual content. 

In the photo generated on Bard, from Google, there are four astronaut dogs embroidered on a round structure.
Image generated in Bard

It is important to highlight that only users over 18 years of age can access Bard. 

How do I generate an image in Bard?

When following the instructions and requesting an image in Portuguese, for example, Bard responds that he currently does not have the ability to develop figures. On the other hand, the artificial intelligence responds by describing the procedure to generate the image, which makes the user return to the same point and receive the same sentence: “I still can't create images”.

The image is a screenshot from Bard, Google's generative AI website.
The system is still unable to create images when the request is made in Portuguese.

But calm down. Remember that Bard, so far, only works in English, so all creation descriptions must be in that language, ok?

But, if you're curious, here are the instructions:

Step 1: open Bard in your browser: www.bard.google.com and log in using your Google account:

Step 2: In the text field, enter the information and details you want to include in the image or photo:

The image is a screenshot from Bard, Google's generative AI website.
The request to create the image must be detailed.

Bard is a developing system and, as it is an AI, it learns with each use. Therefore, use words that capture the essence of the image you want to create. Objectively describe other details, such as colors, shapes, styles, objects, characters and the environment. The more details you provide, the more faithful the image will be to your imagination.

The image is a screenshot from Bard, Google's generative AI website.
Despite asking for a high-resolution image and/or a certain size, Bard still does not provide large images or size options.

The images generated by Bard (in the image above) are square and have a dimension of 1536 x 1536. In other tests, even requesting other sizes, the format is the same.

Step 3: To create more image options, click “generate more”. The option may also appear in English as “generate more”.

The image is a screenshot from Bard, Google's generative AI website.
When requesting more options, Bard will generate new images with the same description that was previously provided.

Step 4: In addition to photos, it is possible to create drawings. To do this, at the beginning of the sentence, always make it clear what style of image you want: drawing; Photograph; crafts or even an image in the style of Van Gogh's arts, for example.

The image is a screenshot from Bard, Google's generative AI website.
With Google AI it is possible to create images as if they were drawings
The image is a screenshot from Bard, Google's generative AI website.
It is also possible to create illustrations using the style of renowned artists

Step 5: To download any of the creations, simply hover the mouse cursor over the chosen image and click on the download symbol.

The image is a screenshot from Bard, Google's generative AI website.
Download is free

The use of AI in image searches

The use of AI in images is something studied by Google for a while now, however, until then the objective was to use photos and other image files to improve searches on the search engine.

In 2011, bigtech launched a search engine using images instead of text. To search for information, the user could upload photos or use the URL of an image to find out, for example, the popular name and species of a plant.

Years later, in 2017, the Google Lens, which in this case can search based on photos captured by the smartphone's camera. The resource is widely used and according to Google, the system currently does more than 12 billion visual searches per month. 

The two systems paved the way for one of Google's most recent releases, the Multisearch, a website that allows you to search using images and text at the same time. According to the company, with Multibuscas you can, for example, take a photo of your set of dishes and then type in the search “coffee table” to find the piece of furniture that matches your dishes.

Source:Google Support, Bard, The Verge e Deepmind Google

See also: 

Bard: understand what Google's AI will do for you

See how to use Bard, Google's new AI chatbot

reviewed by Glaucon Vital in 9 / 12 / 24.

Related Posts