Here Are The Best AI Image Generators

Topline

Here are some of the best AI image generators and how to use them.

Key Facts

Dall-E 2

DALL-E 2 is an AI-powered image generator created by OpenAI, the makers of ChatGPT. The original DALL-E was released in 2021, and DALL-E 2, the updated version, was released in November 2022. Users enter text descriptions into the system, and the software spits out realistic, original images. The Microsoft-backed OpenAI has integrated DALL-E 2 into several platforms, including search engine Bing. It’s integrated into Bing Chat, Microsoft’s AI chatbot, and called the Bing Image Creator. OpenAI has limited DALL-E 2’s ability to create “violent, hate, or adult images,” by removing explicit content from training data. According to OpenAI, users own the images generated by DALL-E 2, meaning they have a right to sell, reprint and merchandise the images.

Artbreeder

Artbreeder is an AI-generated tool run on BigGAN and StyleGAN that allows users to create and modify images of landscapes, faces, paintings and more. It prides itself on its collaborative nature, as users mainly interact by “breeding” (or remixing) other users’ images. To make new variations of an image, users can adjust the image’s sliders, known as its “genes,” which include options like clothing, face, armor and jacket. Publicly shared images can be “crossbred” with each other and adjusted using a slider to determine how much of one parent image the “child” image looks like. Under the collager section, users are able to make their own images by entering prompts that AI generates into pictures. Text-based AI-generated fantasy game, AI Dungeon, uses Artbreeder to create profile pictures for its users. According to Artbreeder’s terms of use, any images created and publicly shared on the site are released into the public domain and can be used “by anyone for any purpose, with or without attribution.”

Let’s Enhance

This AI text-to-image generator was founded by Ukrainian developers in 2017. Let’s Enhance allows users to enter prompts for images and gives several options to change their outcome, like image type, photo style, lens, mood, aesthetic, artist style and lighting. It also is a photo upscaler and enhancer, allowing users to upload their images for correction or enhancement. Some of its capabilities include adding clarity to the image, increasing resolution, adjusting saturation, resizing, light balancing and improving quality. The AI technology removes noise and other imperfections from images, making them look more polished. The company’s founders told PetaPixel the software uses the machine learning frameworks SRGAN and EDSR to upscale images. According to Let’s Enhance’s terms of use, users can use artwork for “any legal purpose,” as long as there’s no third-party content, and it’s in line with the SD License, a contract outlining how and under what conditions images can be used.

Waifu Labs

Waifu Labs uses AI to generate custom anime portraits. Users follow four steps to get their portraits and the AI creates an image based on these preferences. The first step is to choose the initial portrait. Fifteen images of anime-styled characters from the neck up are displayed. If the user doesn’t resonate with any, they have the ability to reset and generate more. Next, users must pick a color palette for the character, which comes in hues of pinks, greens, purples, blues and more. The next step is fine tuning the details. This includes changing the skin color, hair color, eye color, clothing and background. Lastly, users pick the character’s pose. Waifu Labs creates the image and allows users to name the character, download it or use it in the company’s mobile app game, Arrowmancer. The company was created by Sizigi Studios, a group of mathematicians from the Massachusetts Institute of Technology. The company doesn’t disclose whether images can be used commercially—Forbes has reached out for comment.

Deep Dream Generator

Deep Dream Generator’s AI software creates psychedelic, over-processed images that give off a dream-like appearance. Users are given three image processing options. The first is called “text 2 dream,” where users can enter a prompt to generate an image. They can either upload a base image, or start from scratch, and are given options like AI model, face enhance, aspect ratio and upscale and enhance to change the image’s quality. The next processing option is “deep style,” where users first upload their own base photo, then they pick from preselected style images, which includes Starry Night, colorful abstract and Pieter Bruegel. The AI will then spit out the base image in the style of the selected image. The last processing option is called “deep dream.” Users upload an image and the AI then processes it into a dream-like photo, with colorful overlays of animals like frogs, fish and birds. When it comes to using the images commercially, the company only allows it if they were created with a paid subscription or an energy pack, a one-time purchase of credits.

How Do They Work?

According to MIT’s Computer Science and Artificial Intelligence Laboratory, AI image generators are trained on images found on the internet and their patterns and captions. So, if the AI is familiar with pictures of walruses, it’s trained to create images similar to those it’s been trained on when given the prompt “walrus.” The AI is trained to find patterns in art, so it tends to create art pieces “very similar” to what humans have already made. However, Hany Farid, a professor of computer science at the University of California, Berkeley, told PBS the AI isn’t just “cobbling together an image from a bunch of images it’s seen before.” Instead, the online images help the AI learn how to use a text prompt to find a visual noise pattern (random pixels that don’t represent anything) and turn it into a fleshed out image.

Further Reading

Republicans Share An Apocalyptic AI-Powered Attack Ad Against Biden: Here’s How To Spot A Deepfake (Forbes)

Bard Vs. ChatGPT: The Major Difference Between The AI Chat Tools, Explained (Forbes)

Here’s What To Know About OpenAI’s ChatGPT—What It’s Disrupting And How To Use It (Forbes)

You’re Already Using AI: Here’s Where It’s At In Everyday Life, From Facial Recognition To Navigation Apps (Forbes)

Here’s How To Use AI—Like ChatGPT And Bard—For Everyday Tasks Like Creating A Budget, Finding Airfare Or Planning Meals (Forbes)