AI Game Show with SAM

What happens when 300 people co-author a movie trailer plot with AI?

At CIIIC 2026 in Utrecht, Tellart’s AI Game Show turned this question into a live, participatory experience. Guests were invited into a shared encounter with SAM, a custom-built AI character that engages participants individually while curating their contributions into a collective narrative.

Participatory Game Format

Hosted by Tellart Technology Director Sabrina Verhage alongside SAM AI, guests were invited into this “vibe-coded” environment, where collective, AI-mediated decision-making was demonstrated through a game format. The outcome of the experience was a generative film trailer.

Structured across three rounds, participants contributed fragments to a developing story: plotlines, characters, and elements of worldbuilding.

Behind the scenes, SAM operated as a network of custom-built AI agents, each assigned to a single participant. Through personalised prompts, each agent filtered responses individually and selected a few answers.

A team of AI agents processed the accumulated data and translated it into a script, structuring it like a director and composing individual scenes.

From SAM AI’s orchestration to the final narration, sound design, and music, the entire production was generated using AI.

From Interaction to Collective System

The project marks Tellart’s most advanced exploration of AI to date.  Expanding from earlier works such as SAM AI (2024), which established AI as a responsive system, and Sunrise Explorer (2025), which introduced multi-agent structures in a one-person experience, the AI Game Show extends these ideas into a collective format involving 300 participants.

Here, interaction shifted from a one-to-one model to a distributed system. This way, the audience itself became part of the multi-agent system. Their inputs were neither equal nor linear; instead, SAM mediated them into a shared narrative that was ultimately developed into a generative film.

AI Mediation for Experience Design

At Tellart, we approach AI not as a tool, but as a collaborator. AI often produces outputs that differ from what we might expect from a human perspective. By using AI to mediate decision-making processes, we sought to test the boundaries of collective imagination between humans and machines, while also revealing its internal logic as it shapes outputs collaboratively.

While the system successfully orchestrates participation and inputs, the cinematic result often remains “unimpressive” or “ugly,” at times lacking coherence or depth. This gap between process and output suggests that AI, as a collaborator, is currently more effective at aggregating and transforming contributions than at producing refined creative work.

Although the experience draws attention less to the final creative output, it invites reflection on the systems of collaboration and decision-making that underpin many aspects of our social, cultural, and organisational life.