13.11.2023
Kim Schönert about artificial intelligence
WHAT DO YOU DO AT HAJOK DESIGN?
I take photos in HAJOK Design's in-house photo studio and, as an art director, oversee photo shoots at external studios. I do most work in our in-house studio. However, if the job exceeds a certain size, it can make sense to book a larger studio with assistants and/or a food stylist.
WHAT AI APPLICATIONS DO YOU USE THE MOST?
I've used Open AI, Midjourney and Adobe Firefly (generative fill) and I find a combination of Midjourney and Adobe Firefly best suited for our purposes. The basics are done with Midjourney and then specific changes can be made to the created image with Adobe Firefly.
WHAT DO YOU USE MIDJOURNEY FOR? IN WHAT WAY DOES THIS AI-TOOL HELP YOU DO YOUR JOB?
Sometimes certain products are not available due to the season. Here it can make sense to create the product with the help of AI. Even in the first layout process, gaps can be filled with very specific image material without having to work with generic place- holders.
WHAT CHALLENGES DO YOU FACE WORKING WITH MIDJOURNEY?
When using Midjourney, a lot of what is generated is still quite random. You can intervene to a limited extent and sometimes get fairly close to a desired motif by means of the right prompts and parameters, but sometimes it's really hopeless. This is especially the case with more complex content. Image size is still a big limiting factor and simply too small for many of our needs.
WHAT ADVANTAGES OF USING AI EXIST FOR OUR CLIENTS?
Content creation with AI can give a first impression of the final image. Such image material can fill white gaps, especially during the layout phase. To a certain extent, existing images can also be supplemented with AI-generated content. However, in our line of work with food imagery, we will continue to take photos in the future, because we always aim to work with the client's original products. This is also often a requirement of our clients as well as being a legal requirement.
WHAT DO YOU FIND PARTICULARLY EXCITING ABOUT THE POSSIBILITIES OF AI?
The technical background fascinates me. I've had it explained to me several times how it works, but I still don't fully understand it! Luckily, this doesn't hold you back from using AI. And it's a tremendous lot of fun. It’s amazing just how good and also how “creative” some results are and how you can create exciting realities that simply don’t exist.
WHAT STILL HAS TO BE IMPROVED?
Using Midjourney is a bit cumbersome and incon- venient right now. But in the future, it will also work with an own app and website. I'm hoping we’ll get more user-friendly handling. The next step will be a good “inpainting” or “generative fill” feature (both of which are methods to change, flesh out or expand certain areas of an existing image). Under perfect conditions, these tools work quite well, but there is still a lot of room for improvement. Another point is the image size, as already mentioned. We often need higher resolution images. It would also be preferable to have a clear legal framework. Adobe is leading the way here, but I think it does need general rules for everyone.