The capacity to influence the trajectory of AI is concentrated among a very small number of people. The rest of us are positioned as consumers of whatever future they decide to build. We chose the name "OpenOpenAI" to reflect on the paradox of a company whose mission was to democratize AI for humanity while pursuing a closed-source approach. The question we wanted to ask was simple: can we use the very technology under scrutiny, AI itself, to open up public participation in shaping its direction?

The project is a web platform that generates alternate versions of a hypothetical 2024 OpenAI keynote address by Sam Altman. We chose OpenAI's prominence deliberately: it is a topic most people have already formed opinions about, which makes it a productive ground for reflection. The interface presents eight question cards covering topics like the goal of AI, the future of work, AI ownership, creativity, and OpenAI's board composition. Each question uses a range slider rather than a binary choice, because these issues exist on a spectrum. As users adjust the sliders, the system dynamically generates a unique narrative script at the intersection of their positions.

In a pilot study with 30 participants, the platform showed strong results for opening minds to new perspectives and fostering a sense of engagement and control. Users valued the personalization, the ability to see opposing viewpoints without having to seek them out, and the way the experience made abstract AI governance questions feel concrete and immediate. Some found the deepfake element unsettling, which we consider a productive discomfort: it surfaces real concerns about synthetic media while demonstrating the technology's capacity for constructive use.