The reason is that information asymmetry between what generative image AI can and can’t do has only grown in recent months: many still think that ChatGPT is the only way to generate images and that all AI-generated images are wavy AI slop with a piss yellow filter.

If you’re an artist and using AI in whole or in part to create your art, do you think someone “off the street” is going to employ AI and create something better than you, a seasoned pro? Consider for a moment the smartphone camera. “Anyone can take pictures these days, photography isn’t art”. A photographer with a smartphone could surely do better, right? The differences between an average smartphone user and what a professional photographer can do with a smartphone are potentially as vast and wide as the Grand Canyon.

“The fear has sometimes been expressed that photography would in time entirely supersede the art of painting. Some people seem to think that when the process of taking photographs in colors has been perfected and made common enough, the painter will have nothing more to do.” — Henrietta Clopath, 1901

Texture and Pattern Repetition: AI sometimes struggles with complex textures or patterns, leading to noticeable repetition or awkward transitions. Students should look for unnatural patterns in textures like hair, skin, clothing, or background elements.

A study by Sung and Mayer (2012) suggests that any graphic in a learning experience will fall into one of these three categories: Instructive images: These visuals directly support learning and facilitate essential cognitive processing of core concepts. For example, a diagram illustrating Porter’s Five Forces can help students better understand this business strategy framework. Decorative images: These graphics enhance aesthetics but don’t influence learning. For example, an image of a business handshake can be visually appealing but won’t support or obstruct students’ understanding of negotiation strategies. Distracting images: Sung and Mayer call this category “seductive” images. While these visuals may relate to the topic, they impede learning because they require extraneous cognitive processing. As an example, consider a complex organizational chart of a full corporation in a lesson on team leadership. The image connects broadly to the lesson but also highlights a lot of irrelevant details, distracting students from the key concepts.

While LLMs generate text one word at a time, always moving forward, diffusion models start with random static and transform the entire image simultaneously through dozens of steps. It is like the difference between writing a story sentence by sentence versus starting with a marble block and gradually sculpting it into a statue, every part of the image is being refined at once, not built up sequentially.

But what makes diffusion models interesting is not their increasing ability to make photorealistic images, but rather the fact that they can create images in various styles.

Unlike diffusion models that transform noise into images, multimodal generation lets Large Language Models directly create images by adding tiny patches of color one after another, just as they add words one after another. This gives AIs deep control over the images it creates.

First, we prompt seven state-of-the-art generative text-to-image models (DALL-E 3, Midjourney v6, Firefly 2, Imagen 2, Imagine, Realistic Vision, and Stable Diffusion XL Turbo) to create 10,320 synthetic marketing images, using 2,400 real-world, human-made images as input. 254,400 human evaluations of these images show that AI-generated marketing imagery can surpass human-made images in quality, realism, and aesthetics. Second, we give identical creative briefings to commissioned human freelancers and the AI models, showing that the best synthetic images also excel in ad creativity, ad attitudes, and prompt following.