Last week, artificial intelligence-generated photos depicting actors Deepika Padukone and Hrithik Roshan as older people went viral. Many users have been sharing similar pictures of celebrities and influential people from across the world.
In recent months, artificial intelligence (AI) art has taken over the Internet and the launch of text-to-image generators has taken over people's imagination. Now, it’s easier than ever to see the world through art; from imagining living with Kermit the frog as painted by Munch to making the Cookie Monster part of Burlesque (2010).
Some Indian creators have made creative use of AI to show a different world. Visual artist Cezanne Ali and video creator Bijoy Shetty used AI art to imagine a hypothetical, sci-fi world where India was cohabitated by aliens, while artist Sahid posted AI-generated pictures of metros such as Bengaluru and Hyderabad blending their cultures, food, and development.
With more people turning to AI art generators to give life to their imaginations, we look at some easy ones to get started with.
Last year when OpenAI’s chatbot wowed the Internet with its AI chatbot, ChatGPT, it also released DALL-E 2 to turn simple words into intricate artworks. The first version, DALL-E, released in January 2021, is a 12-billion parameter version of GPT-3 (a large-language model) trained to generate images from text descriptions, using text–image pairs.
Let’s look at a basic example. If you are looking for a picture of a chair shaped like a strawberry, you have to type just that and Dall-E will create a realistic-looking picture. The newer version of the generator, DALL-E 2, creates more realistic and accurate images from text description with four times greater resolution. It combines concepts, attributes, and styles and OpenAI hopes it will empower people to express themselves creatively. While users own the images, they create with DALL-E, they have to credit DALL-E 2 using a watermark.
You can create photos on DALL-E 2 using text prompts, which refer to image descriptions, variations, created by uploading existing or generated pictures, or edits wherein you can delete or make changes to features of an image.
It also has a feature called outpainting, which enables you to expand the landscape of the images, which helps adjust the size of the image. From creating hyper-realistic images to using AI-generated product images for marketing, people are making use of DALL-E in interesting ways.
If you have seen images of the Pope wearing a Balenciaga puffer or former US president Donald Trump getting arrested that went viral earlier this year, then you are familiar with Midjourney.
One of DALL-E’s biggest competitors, Midjourney has a unique style of generating AI images. While it can create realistic images, Midjourney is well-used to add a dream-like essence to the pictures. According to the official website, this AI art generator aims to “explore new mediums of thought and expand the imaginative powers of the human species”.
Midjourney has been consistently updated since it was first launched in March 2022. Its fifth version was rolled out in March this year. Currently, you can use Midjourney only through its official Discord channel.
To generate images, you can use the “/imagine” command to enter the prompt. Four options will be shown and you get to choose which you want to move ahead with. You can also use the img2img function to upload an existing image to the Discord server and use it as a template to generate new images.
Founder David Holtz announced in March that Midjourney is ending free trials citing “extraordinary demand and trial abuse”, according to a report in The Verge.
Formerly known as DALL-E mini, Craiyon is designed by developer Boris Dayma as a free text-to-image AI tool. To generate images, you have to type a description, just like other AI art generators. Interestingly, you can also add negative words to ensure the items you want to exclude from the generated results. Finally, choose if you want an art, drawing, or photo.
You can choose to add art style descriptions such as Van Gogh or Impressionism, composition descriptions such as aerial view and mood descriptions such as vintage or futuristic, according to the Lab Lab AI website.
Another free and easily accessible AI Art generator, Fotor, gives users the option to create diverse images using prompts; from science fiction to 3D. Fotor uses the Stable Diffusion AI art generator model and includes many options such as an AI face generator, to create hyper-realistic faces based on traits such as gender and age.
You can create photorealistic, illustration, cartoon, anime and pixel art using this tool. There is also an option for customizing the AI artworks by fine-tuning settings such as art styles, dimensions, compositions, shadows, and highlights.
One of the widely used tools by graphic designers, Canva, has now also entered the AI space. With the Canva AI art generator, you can visualize a product or idea, bringing your imagination to life using text prompts.
When you can’t find the perfect image, all you have to do is think of the right words. Canva’s AI art generator creates unique images from text descriptions that can be used for social media posts or presentations. “Text to Image unlocks creativity in all sorts of ways and empowers all sorts of people, from content creators to marketers, professional designers, and social media managers,” Canva said in a blog post announcing the AI art generator in November 2022.
Today, most professions need to produce content quickly and make sure it stands out. In such an environment, having a tool like this can be a big help.
Canva also has a Magic Write feature, powered by OpenAI, to create a first draft based on your idea. From social media captions to profile bios, this can help you brainstorm ideas for the perfect social media post and unlock writer’s block.