What happens when you ask artificial intelligence to imagine the way new products will look?
What happens when you ask artificial intelligence to imagine the way new products will look?
Technology reporter Mike Hasabella asked the DALLE-2 AI engine to generate images for him of the way household appliances would look if Tesla made them. The results are in front of you: in a bright red color and with a combination of neon lights and a futuristic look. That is, except for the robot, which looks like it was taken from a futuristic movie from the XNUMXs, with the addition of color.[1]
You can love or hate the ideas of artificial intelligence, but it is important to put your finger on three points for the future.
First, we see here in real time how the work of the artist and graphic designer is changing and requires different skills than those that have been accepted until now. What skills? according to the assembly -
"It's worth noting that inserting text titles is an art in itself, and there are so many techniques for using language to create images."
The DALLE engine generates images based on titles that the artist gives it. The titles can be simple and short, like "washing machine", or much more in-depth, like - "A washing machine with a futuristic scent, made by Tesla, in a realistic style from a high angle of photography, with the photographer's reflection on the right."
In other words, even using DALLE requires certain skills. The people who will get the most out of the artificial intelligence will be those who understand the theory behind the different ways of photography and who know how to use the correct and most accurate words to get the best result on the first try. They will have to know how to visualize their visions not only in diagrams but also in words. And of course, they will have to understand how the artificial intelligence 'thinks' - at least at a sufficient level to direct it in ways that are not always obvious to those who have not gained experience working with it.
In a certain sense, the textual titles are the ones that become the real art, or at least an inseparable part of it. And although it's probably impossible to patent or copyright them, some artisans will make them a trade secret. Even Assabella, who is not an artist by profession, announced that - "I kept the titles that I used for the benefit of the works to myself, but I chose to share the results with you."
Second, we only see one image here for each "title" that Isabella fed to the AI. If you don't like the same picture, don't worry. The artificial intelligence will generate dozens or hundreds of new options for you, until it reaches the image of the most suitable product for you. Such is the future: with endless possibilities for vacuum cleaners of all shapes. The role of the craftsman is mainly to choose from the abundance the option that will appeal to the largest number of possible customers.
Thirdly, the future will allow us to enjoy an abundance of products of all kinds. A significant part of the cost of producing a new product today comes from the human design work. Using XNUMXD printers and other sophisticated robots, companies will already in the coming decades be able to make adjustments to many products so that they look as the customers request. We will not have to settle for vacuum cleaners in different colors, but we can also get them with wheels (or without), with horns (or without), and with a variety of other features and shapes as we wish. This development will not happen immediately, as it is much more difficult to move from standard robotic production lines, to flexible robots with different capabilities. And yet, we are witnessing here the first steps in the direction.
Good luck in the future, which will be full of more colorful, varied and interesting household items!
Or they'll all be bright red, if we let Tesla design them.
[1] https://medium.datadriveninvestor.com/dall-e-2-imagines-tesla-household-products-df2fae7254df
More of the topic in Hayadan:
