Not long ago, I was doing research for a film about ice cream. Not knowing exactly how many flavors the brand had, I tried asking ChatGPT, which not only provided me with an exact table of all the flavors but also generated an image showing the layers of the double chocolate flavor to better explain how it’s made. The fact that it chose to generate a custom image for me rather than simply searching the internet for a real product photo makes one reflect on how easy it has become to generate images almost instantaneously with artificial intelligence. There are numerous tools: from ChatGPT itself to the highly popular Midjourney and DALL-E, and many more.
Physics in the AI world
For all creative agencies, the use of artificial intelligence has become as fundamental as the internet. It not only assists creatives in writing stories but also allows clients to visualize a scene before it’s even created, with appealing and engaging images. However, this significantly raises expectations for the final results, expectations that must be met or even exceeded. Compared to a few years ago, when a simple drawing was used to illustrate a scene and the skill lay in the agency’s account and creatives convincing the client that this was the right path, today it’s a whole different playing field. The client immediately gets a visually engaging result that creates expectations, ideas, and thoughts.
Table top, on the other hand, takes the viewer into an extraordinary world, where action is seen in slow motion, especially at 1000 frames per second, capturing details that would normally be invisible. It’s a way to show reality in an amplified and spectacular form, yet still real. Now, with AI, the playing field has completely changed. In creatives’ visuals, there are often situations that are impossible to achieve, images that blend fish-eye and telephoto lenses, the physics represented often doesn’t exist, but the image remains appealing due to the color combinations, framing, and textures. When I saw Sora’s work, I was astonished: it’s almost as if physics has made a quantum leap, and the physics here seems to work. Suddenly, Sora has outshone all other competitors on the market, and as soon as it becomes available to the public, I’m sure we’ll see great things.
From imagination to reality: the role of professionals
No matter how advanced AI becomes, the realization of these ideas remains in the hands of industry professionals. Directors, cinematographers, food stylists, and many others collaborate to turn digital visions into reality. The challenge is to take the perfect, sometimes surreal images created by AI and translate them into real shoots, maintaining the essence and originality of the initial concept.
For instance, in the web and TV campaign we created for SMEG, we used gigantic translights generated with Midjourney and then upscaled with Magnific.ai, another software that enlarges images and adds details. We needed many locations, and conducting an iconographic search to find images with the desired lighting would have taken much more time. Instead, by leveraging artificial intelligence, we were able to seamlessly blend the scenographies created with Midjourney with real-life shoots, getting the most out of AI integrated with our creative work.
