“Artificial intelligence is both terribly interesting and dangerous”

All summer long, we interview employees, freelancers, and business leaders about their relationship with generative artificial intelligence. How do they use it, how do it change their professional practices? Today, Anne Fraysse, a freelance graphic designer in Lyon.

In graphic design, Anne Fraysse specializes in brand design. That is to say, she designs the visual identity of a company, which allows customers to recognize it at a glance. She says she became interested in generative artificial intelligence, with the arrival of services like Midjourney and Dall-E, capable of generating images.

“As a designer, our profession evolves according to the tools, always in a quite spectacular way. I am 51 years old, when I started working, we made models with photocopies, we did not print in color. What changed our work was the Internet. It is the same with artificial intelligence: they will fundamentally change our practice.”

After testing generative AI for a year, Anne Fraysse has been using it regularly for several months.

“Artificial Intelligence allows me to easily create paradoxical images and explore ideas in a very free way. For example, when I worked on the visual identity of a communication agency, I gave Midjourney an image of a renaissance painting and asked it to add high-tech products, an iPad and a headset. Midjourney generated an image with a renaissance engraving graphic style, quite astonishing, quite disturbing and therefore interesting. It would have taken me days to do it.”

Generative AI is a time saver and an aid to creativity. That’s how I want to think about it.

Anne Fraysse, graphic designer

to franceinfo

Today, the graphic designer defines these generative AIs as “terribly interesting and dangerous” at the same time. She points out a risk of standardization of creation, if the whole profession starts using them. But what she fears above all is losing customers.

“There are platforms that allow you to generate logotypes that are, for the moment, quite bad but which already satisfy a certain number of people. It is quite possible that in 4 to 5 years, more efficient platforms will emerge and that they will be able to produce perfectly correct visual identities for 90% of my clients. This worries me.”

The graphic designer also raises the question of copyright and the remuneration of graphic designers whose work is used to feed and train these generative AIs.