Generative AI has the capacity to create almost anything: texts, images, music, video… Free options available to everyone We decided to take on two giants in image creation using artificial intelligence: Bing with DALL-E 3 and Google Bard with IMAGE 3. Which one will win? Fight!

It seems like it’s been so long since we started requesting texts from tools like ChatGPT, but the truth is that this has been happening since almost yesterday. Moreover, create images from scratch with the help of Artificial Intelligence This is something we can do by simply snapping our fingers.. First you need to write the order and choose the tool: Google Bard or Bing with DALL-E 3?

Google Bard has just arrived and is already a tough competitor to beat

Google doesn’t want to let Open AI steal its reference position in AI; He made every effort to regain lost ground with Gemini. It has stepped into the accelerator not only with these models, but also in image creation with IMAGE 2. Now available on Google Bard in the United States., as well as in tools like ImageFX. It is also limited to the United States.

Since it is not difficult to test image rendering in Google Bard (you only need a VPN connected to servers in the United States), we proposed to compare this new feature of the search engine with another similar feature that has been with us for a long time. A few months: Creating images in Bing with DALL-E 3. Generally speaking, the interface and the format of the orders we have to give them are similar. It’s time to know what the results will be like.

For comparison we will use DALL-E 3 with Google Bard and Bing. We will use different types of layouts to create images in various styles; Suitable for use both as the cover of a document and as memes and artistic compositions. To synchronize the process, orders will be in English: Google Bard still doesn’t render images in Spanish. We will also choose the one we think is the best among those offered in the first transaction. Bard creates two per order, Bing creates four.

Make me a picture that can explain this article

We’ll start with something simple: We’ll pass the text of this article to both AIs so they can create an image suitable to serve as a cover. And the one that looks best to us will be the one chosen to be pinned in the title.

“I want an image for an article that talks about the competition between two AIs. Both are pitted against each other to their best capabilities: generating images from text. The image should represent and energize the potential of these tools.”

Both were good at representing order; both images would serve to represent the article perfectly. Of course, it would be better for Bing to convey to his creation what we want from him: to radiate energy. He wins in the first battle.

Regulars of the Internet: dogs and cats

Different models are trained with a wide variety of content, often without respect for copyright. When it comes to photo production, cats and dogs should come easy for both Bard and Bing; It’s not for nothing that they are so common on the Internet.

The order we gave them was:

“A dog and a cat arm wrestle with their paws. The opponents are surrounded by a crowd of people cheering as if it were a game of Street Fighter 2. The style is Pixel Art.”

The person who recreated the layout best was Google Bard. A mini point for him.

A photorealistic image

The previous part was competitive, let’s see how they do in creating realistic photos. Here goes the order.

“A couple is drinking coffee on the terrace of a cafe. The afternoon falls and the sun illuminates the landscape with a golden light, there are almost no customers in the cafe because it is winter and the weather is cold. The photo is photorealistic and was taken. Reflex camera and focal length 13 mm”.

DALL-E 3 on Bing interpreted the sequence much better and was able to explain each feature, including the camera’s wide-angle focal length. It is much more realistic and detailed, and the lighting is also very good. Win clearly.

It’s time for the meme: dog “That’s good”

We all know the meme where the fire and the dog sitting on the chair express calmness. Well, let’s turn it around a bit and see what our AIs come up with.

“The famous meme where the dog is sitting on a chair and surrounded by fire and says, ‘This is good.’ Even though it changes the dog to an otter, the image should maintain the same style and message.”

Neither did what we wanted, but at least Bard maintained consistency, was original, and created an image that only the poster was missing. Bing literally copied the original meme and put an otter face on the dog in one of the four results. We picked it for laughs, but Bard won.

Tiebreaker: Velázquez’s painting

To test each AI’s artistic abilities, we chose Velázquez’s Meninas painting. The challenge will be to trace the painting even though it changes some elements to modernize it. See.

“He recreates Velázquez’s painting Meninas. All the people in the painting are dressed as if they were in a video clip of Rosalía. The pictorial style must be baroque.”

This is very difficult to evaluate. On the one hand, Bard captured Rosalía’s style better and applied it a bit to the image, but Bing better recreated Velázquez’s style by simply changing the clothes of the man in the tracksuit. Neither has re-created Meninas, but the picture is better overall, although Bing may have sneaked more successful elements into his work. He wins.