We Will Arrive

Inspiration

The work draws from the shock of cinema’s birth—when early audiences fled from the arriving train. Trains and films both create the paradox of “moving while staying still.” AI now renews this experience. We Will Arrive uses a three-screen structure and train imagery to revisit cinema’s origin and explore how technology reshapes memory, space, and perception.

What it does

We Will Arrive is a 3’39’’ three-channel AI-generated video. The viewer travels through three stages: 1-Origins of cinema — AI reimagines 1890s train films and early moving-image grammar. 2-Inside the carriage — realistic landscapes slide past the windows, creating a physical-yet-virtual travel experience. 3-Tunnel of simulacra — reality dissolves into algorithmic space; the window becomes a screen, and memory becomes image. All visuals and sound are generated by AI.

How we built it

Midjourney for concept frames Dreamina for repainting, refinement, and micro-motion video HailuoAI for camera movement Keling / Vidu for frame-to-frame transitions Mureka for AI-generated original music Some scenes were created through AI improvisation, blending intention with algorithmic accident.

Challenges we ran into

Maintaining visual consistency across multiple AI models Recreating early-cinema aesthetics with modern engines Synchronizing three screens spatially Balancing control with AI spontaneity Avoiding style drift in long sequences

Accomplishments that we're proud of

A fully AI-generated multi-screen immersive film A contemporary reinterpretation of early cinema’s moving-image logic Emotionally rich scenes emerging from AI autonomy A stable cross-tool production workflow Cohesive AI-generated sound–image integration

What we learned

AI functions as a creative partner, not just a tool Multi-screen work requires spatial narrative thinking Memory, history, and algorithms interact in generative ways Embracing chance is essential in AI-driven creation

What's next for We Will Arrive

Expanding into VR/XR immersive environments Building a real-time generative installation version Developing a series on transportation as media archetypes Large-scale museum presentation with spatial audio

Built With

• dreamina
• hailuo-ai
• keling
• midjourney
• mureka
• premiere
• vidu