What does “a pink guinea pig choosing a gala dress from a wardrobe full of sharks” look like according to an AI? Lately, several AI art generators have challenged internet users to come up with the most unexpected descriptions to create fictitious, sometimes extremely bizarre, images using artificial intelligence. But how does the technology work? And what could it be used for in the future?
There are several ways to use AI to create images. One approach is to train computers to recognize patterns in data using deep learning algorithms. Once trained, the algorithm can produce similar results when presented with new data. Another approach uses generative adversarial networks (GANs), in which two distinct models compete with each other. A third technique uses reinforcement learning, where the system learns by trial and error.
Two applications that have recently gained media attention are DALL-E and Midjourney DALL-E developed by OpenAI uses a modified GTP-3 version and a deep learning model to generate images. Midjourney was developed by a small research team and similarly uses neural networks to generate images. Both apps are in open beta for private users. If you want to try DALL-E, you can register for access at this link.
How can this technology be relevant for the fashion industry?
Generating images based on large amounts of data has the potential to become a valuable technology trending in various fields. Stock photo services may be replaced by artificial intelligence applications in the future as the technology becomes more refined and precise.
AI generators have the ability to take a large number of images to create a realistic new look while stretching the human imagination beyond its limits and biases. If a product designer wants to create a large number of iterations of a prototype, for example, an AI image generator can deliver that in minutes or even seconds. In this article, you can see an example of how an AI powered by shoe image data can be trained to create new out-of-the-box designs.
Enabling a lot of data that might be difficult for the human eye to see quickly can be used in trend analysis and forecasting when numbers and statistics don’t provide a quick or good enough overview. The production of photos and images will probably in the future be partially replaced by images generated by AI or other technologies and tasks that complement manual work.
The film, gaming and digital entertainment industries will be able to use similar applications to create characters or environments. In the current state of open betas, the release can now be used as concept art and inspirational input where, in the future, it is easy to imagine an AI in a more rigorous and refined framework that creates assets ready for the production. Being able to quickly iterate and modify images through manual feedback goes far beyond what humans can achieve in the same amount of time.