Co-Creating with Machines - An Exploration of Generative AI

Co-Creating with Machines - An Exploration of Generative AI

As someone deeply passionate about processes of co-creation, I’ve spent years working with methodologies that bring people together to unlock collective intelligence. Whether it’s design thinking, metadesign, or sprints, these frameworks tap into a form of real intelligence, one rooted in the lived experiences of individuals who come together to share, reflect, and innovate. This is the kind of intelligence that I find most fascinating — collective and reflective.

However, as much as I cherish these human-centered processes, I’ve also become an enthusiast for another kind of “intelligence”: generative AI. Unlike the collective intelligence born from real-world experiences, generative AI is built on vast datasets, learning patterns from everything that has been written, said, or created. This form of intelligence is not reflective or experiential; it’s a model of language, not of the world. Yet, it holds incredible potential when integrated into the creative process.

Bridging Collective and Generative Intelligence

In my work, I’ve been exploring how these two realms of intelligence—collective and generative—can come together in a meaningful way. Generative AI, particularly models like ChatGPT, excels at providing insights, suggesting ideas, and generating content based on patterns it has learned from massive amounts of data. But there’s a key difference: AI lacks lived experience.

For me, co-creation isn’t just about generating ideas—it’s about the dialogue, the shared experiences, the stories that people bring to the table. These are the building blocks of what I call collective intelligence, where participants contribute not just knowledge but the wisdom that comes from their unique life experiences. This is something AI can never fully replicate, no matter how sophisticated its language models become.

The Four Mastermind Minds: A Dialogue with AI

When I’m in a situation where co-creation with others isn’t possible—when I’m working remotely or don’t have collaborators at hand—I often turn to what I call the [[Four mastermind minds]] approach. It’s a prompt I use with ChatGPT, where I ask the AI to create four distinct characters, each with their own name, background, and personality. These virtual “minds” serve as my sounding board, helping me brainstorm ideas, generate content, or just keep the creative juices flowing.

This technique has been invaluable for me, especially when I’m feeling stuck. The process of dialoguing with these four virtual personas gives me a sense of interaction, even when I’m working solo. It’s like having a brainstorming session with a team—except the team is made up of generative AI characters. The AI doesn’t just generate ideas; it helps me unlock new perspectives by engaging me in a sort of creative play.

Where Generative AI Falls Short

But here’s where things get tricky: no matter how well-crafted these virtual minds are, they are not real people. They don’t have real experiences to draw from. They can’t share the stories of their lives, reflect on their pasts, or bring the kind of deep, experiential insight that comes from living through challenges and triumphs.

For example, when I ask ChatGPT to create four characters like “Sofia” and “Natan,” it can generate detailed profiles, give them backstories, and even simulate dialogue between them. But these characters lack the lived experiences that make human co-creation so rich and meaningful. There’s only so much I can build with them before I hit a wall—a point where the lack of real-world experience becomes a limitation.

Next Steps: Bringing Generative AI Deeper into Co-Creation

Generative AI tools, like ChatGPT, excel at specific tasks—they’re fantastic for summarizing, analyzing, and formatting content across various mediums. Whether it’s taking a transcript from a meeting and turning it into an email or repurposing the same content as a blog post, the ability to format and transform is one of the AI’s strongest capabilities. I believe that as these technologies evolve, they’ll become more than just assistant tools; they’ll emerge as true co-pilots in creative processes.

In a co-creation setting, where ideas are generated rapidly and group dynamics are at play, there’s an interesting opportunity to incorporate generative AI as an active participant. Imagine having an AI that listens during brainstorming sessions—where human participants are fully immersed in verbal communication, using not just words but body language and emotion—and helps by structuring, analyzing, or highlighting key topics. The final decisions, of course, would always remain with the humans, but the AI could serve as a facilitator, streamlining the flow of information and insights.

I’m particularly excited about the prospect of designing a generative AI that not only listens but also speaks and contributes in a more dynamic, real-time way during these processes. There’s something unique about the immediacy of spoken dialogue in co-creation; it’s fast-paced, reactive, and often taps into emotional and sensory communication that written text can’t capture. Building an AI that can truly engage in these moments, not as a passive assistant, but as an active agent in the creative process, is an experiment I’m eager to pursue.

Much like I’ve experimented with my “four mastermind minds” prompt, where AI-generated personas like Sofia and Natan help me when I need remote, dialogic input, I see potential in creating AI personas with specific areas of expertise. For example, an AI specialized in narrative building or engineering could join the team as an equal contributor. These personas would still lack the lived experience that makes human intelligence so powerful, but their ability to navigate and synthesize information rapidly could make them valuable partners in creative work.

Ultimately, I don’t see generative AI as just another assistant; I think there’s room for it to become a second brain, deeply integrated into the creative process. It could help organize and enhance our thinking, not just act on orders. While I’m still exploring what this would look like in a real-world setting, it’s a future I’m excited to experiment with. I’m not sure yet which of my partners might be interested in such a project, but I believe this is a space worth diving into.