There are 750 billion images on the internet, and pictures paint a thousand words.

As a writer with miserable artistic skills, sometimes I go speechless.

I wanted to create a vision of an Indian man wearing a plain colored turban in a rainbow-colored suit. My descriptive narrations would not keep my 7-year-old’s concentration for 5 seconds, and google could not produce any results!

I wanted to add colors to the typical dry and monotonous suits preferred in the west over a man in a turban from the east and show her that colorful clothing is part of the culture.

Quite specific!

So we stepped into the world of Artificial Intelligence!

What Is Artificial Intelligence (AI)?

Artificial Intelligence has improved significantly over the past 5 years. AI makes it possible for machines to learn from new inputs and perform human-like tasks and sometimes aim to do things that humans cannot, like converting text to images.

With the help of technologies such as Stable Diffusion and DALL-E, we created strikingly realistic images from just a description!

Stable Diffusion

Stable Diffusion is an art-text-to-image model that generates images from text.

Description: ‘An Indian man with a turban in a rainbow-colored suit’.

Stable Diffusion merges multiple pictures into a hyper-realistic photo. When an image is created from text, it looks like it’s a photograph.

How To Use It?

This is the easiest part, type the description and generate the image.

How about a purple horse galloping on the moon? 

How Does it Work? 

Stable Diffusion determines the elements of the text and merges a few images together to create a new one. Images come from a dataset, almost like a library of images with the corresponding text. 

When the system is given a sentence, using its AI feature, it combines pictures that seem relevant. It does not create a drawing, but a hyper-realistic photo. 

Creators of Stable Diffusion are a part of the growing ecosystem working on image recognition and image manipulation. 

Is The System Safe To Use?

Stability Diffusion, Dall-E, and other creators are still in their research stages. Direct keywords related to nudity and violence and even celebrities or public figures are filtered or deemed unsafe. However, once in a while, a disturbing image can be created.

 For example: ‘Baby playing with a big toy gun’

Those eyes are scarier than the fact that the baby is holding a gun!

Systems like these can potentially be used to target any individual and cause unintentional harm.

 DALL-E 2, is one of the best Ai art generators in the market. it is moving ahead in stages by inviting users via a waitlist to create a diverse and useful experience as they gather data and develop their systems. DALL-E 2 has an easy user interface and creates images of diverse backgrounds when texts are applied. This is an exciting idea when a global audience is set to experience digital systems on a large scale in professional, personal, and creative ways. 

Example of Art generated with DALL-E 2 with similar text used in Stability Diffusion.

What Is The Future Like?

Using AI to create images is going to help creatives and marketing businesses. However, with experimentation, artists may soon do away with professional image manipulation software to create artwork. AI may soon even make Photoshop defunct.

The possibilities of AI technologies are endless but come with their own set of dangers. It is up to us to gain knowledge, so we can impart our wise experiences to the coming generations.

An example of creating something I couldn’t find on the internet.