In April, I found myself in the middle of a creative whirlwind with the team at Mysteryland. What started as a training session about applying generative AI in the creative process turned into a race against the clock to develop the final key visual for a marketing campaign.

What followed was a surprising demonstration of AI’s potential, resulting in a unique image now visible all over Amsterdam in Mysteryland's latest outdoor campaign for 2024.

‍Generating consistent visuals‍

The world of generative AI is evolving at breathtaking speed. For me, image generation is one of the most fascinating aspects of this revolution: creating key visuals based purely on text.

In the summer of 2022, Roy (founder of y=f(x)) gave me access to DALL-E 2. Roy had purchased fifty credits and simply said, "You should experiment with this." I had no idea what to expect, so, like many others, I typed my first prompt: "Sketch Main Stage technology festival lights video crowd sitting led thenextweb." Suddenly, four images appeared from noise before my eyes. Here’s what emerged:

These initial images were still very basic, but the idea that these rough sketches arose from a few typed words was fascinating. And even though the generated images weren't yet commercially usable, it inspired me to keep experimenting. After all, generative AI is an exponential technology—surely it wouldn't be long before newer models bridged the gap to hyper-realistic images suitable for professional use. Just twelve months later, I had my answer.

‍From words to visuals — how does that work?‍

Just as language models (LLMs) predict words to form sentences, image-generation models predict pixels to create images.

This technology is known as diffusion. As with language models, many large and small companies compete for market share with their own diffusion models. The most relevant players currently include DALL-E (from OpenAI), Midjourney, Firefly (from Adobe), Ideogram, and Leonardo.
‍

In the world of image generation, we're all participating in a camel race, much like those found at carnival games. Each new update of these underlying models brings a wave of innovations, each tool managing to outperform the others through leaps in quality, new functionalities, or more refined interpretations of user instructions. The progress is mind-blowing!
‍

If you asked me right now which tool I prefer most, my choice would undoubtedly be Midjourney. This tool is known for its unmatched quality and offers the greatest control when generating unique visuals.‍

Democratization of visualization‍

Why is this revolutionary? The power of image generation lies in its ability to quickly and efficiently produce high-quality visuals — something that previously took hours or even days to create manually. This opens up a world of possibilities for event professionals, from designing visual concepts, mood boards, and stage designs to creating personalized marketing materials.

‍Collective creativity‍

Moreover, image generation changes the way we brainstorm ideas and stimulates collective creativity. Within a group, we can now instantly transform ideas into hundreds of visuals, dramatically accelerating how quickly we arrive at that one perfect concept.

‍Practical applications for Event Professionals‍

As an event professional myself, I've discovered you can use image generation in numerous practical ways, including:

• Sketches: Immediately show your client the core of your concept through rapid sketches capturing the essence of your ideas.
• Key Visuals: Capture the essence of your event in compelling visuals, bringing your story vividly to life.
• Storyboards: Generate your event scene-by-scene. Visualize every detail, clearly communicating to your client exactly what will happen.
  ‍

• Stage Designs: What was once a time-consuming process becomes reality in an instant. From idea to design in the blink of an eye.
• Signage: Create directional signage and informational signs that are functional, feasible, and artistic.
• Branding: Develop a cohesive visual identity for your event, from striking logos and color palettes to style elements and merchandise.
  

• Personas: Visualize your ideal attendee in vivid detail, enabling your team to quickly address their needs and expectations.
• Pitch Documents: Give your proposals a visual edge, seamlessly combining style and substance.
• Content: Effortlessly create visual content that strengthens your brand, from engaging social media images to professional holding slides.

As you may have noticed by now, the possibilities are endless.

In the coming year, we'll dive even deeper together into the exciting world of image generation. Prepare for an inspiring journey through the latest developments, real-life examples, and concrete use cases demonstrating how these powerful tools can be leveraged. We’ll not only explore the technology behind the magic but also share practical tips and tricks to fully harness image generation. So, hold onto your hats! Oh, and one more thing…

You might be curious about the current Mysteryland campaign visible throughout Amsterdam. Here’s the final result of human collective creativity, enhanced with a small touch of Generative AI: