Google DeepMind collaborated with director Eliza McNitt and Darren Aronofsky's company Primordial Soup to produce ANCESTRA, a short film visualizing McNitt's personal birth story using Google's Veo generative video model alongside traditional live-action footage.

    Creative Idea

    Eliza McNitt used Google's Veo model and live action to visualize her birth story.

    Director Eliza McNitt partnered with Google DeepMind and Primordial Soup to create ANCESTRA, a personal short film using the Veo generative model to blend live-action footage with visual sequences of microscopic biological processes and cosmic phenomena.

    Prompting as Poetry and Cosmic Birth Storytelling

    Bringing Poetic Prompts to AI Cinema

    Written and directed by Eliza McNitt, *ANCESTRA* translated her personal story of being born with a hole in her heart into a cinematic narrative depicting an emergency delivery where a mother draws on cosmic forces to save her newborn. To direct Google DeepMind's Veo generative video model, the creative team from Google Creative Lab alongside McNitt rejected standard technical language in favor of a method they termed *"The Poetry of the Prompt."* They fed the AI poetic, emotional language paired with McNitt's actual personal medical and photographic archives from her birth to guide the generative visual output.

    Blending SAG-AFTRA Actors with Generative Twins

    The project established a hybrid workflow involving over 200 traditional film production experts, including live-action crews, SAG-AFTRA actors starring Audrey Corsa, Max Gordon-Moore, and Melanie Nicholls-King, alongside Google DeepMind researchers. Visual effects supervisor Aaron Raff and senior research scientist Kory Mathewson led technical efforts to augment filmed performances. By using Veo to generate dynamic baby sequences, cosmic phenomena, and microscopic cellular activity, the production bypassed the physical limits of working with live infant actors on set, treating prompt adjustments similarly to camera setups.

    Collaboration and Sound Craft

    Darren Aronofsky served as executive producer via his venture Primordial Soup, mentoring the project alongside executive producers Demis Hassabis and Robert Wong. Speaking on the evolving workflow, Aronofsky noted, *"Filmmaking has always been driven by technology... sound, color, VFX allowed us to tell stories in ways that couldn't be told before. Today is no different."* The short film featured a score composed by Caroline Shaw and Danni Lee Parpan, performed by Ringdown, with sound design by Q Department & Mach1. Following its premiere at the 2025 Tribeca Festival, the production methodology became a primary case study at the Sundance Institute's Story Forum.

    4C Framework deconstruction

    Company

    Google DeepMind provided its frontier generative video model, Veo, capable of generating imagery from poetic prompts and archival source material.

    Category

    Generative AI demonstrations often rely on text prompts to render synthetic, standalone video clips without narrative cohesion or human connection.

    Customer

    Cinephiles and filmmakers value deeply personal stories, visual authenticity, and craft over novel technological demonstrations.

    Culture

    The growing debate over artificial intelligence in media spurred interest in how filmmakers might integrate AI into traditional film workflows.

    Strategy:

    Merge generative AI technology with personal live-action filmmaking to explore new avenues for cinematic storytelling.

    How the 4C framework works

    Strategy Technique

    Borrow Equity

    Google DeepMind partnered with acclaimed filmmaker Darren Aronofsky's production venture, Primordial Soup, to anchor its experimental generative AI technology within established cinematic storytelling traditions.

    Explore Technique

    Creative Technique

    Cutting-edge Tech

    The short film utilizes Google DeepMind's Veo AI video generation model to visualize complex biological and cosmic phenomena that cannot be captured with traditional camera setups.

    Explore Technique

    Craft Breakdown

    ANCESTRA excels by pairing live-action SAG-AFTRA performances with custom generative video outputs to depict complex cosmic and anatomical concepts.

    AIExceptional

    Veo generated continuous sequences of microscopic cellular processes and cosmic visuals guided by artistic prompt poetry and personal archival photos.

    Visual EffectsExceptional

    VFX supervisor Aaron Raff seamlessly integrated generative AI video renders into a traditional live-action narrative timeline.

    Cinematography

    Minka Farthing-Kohl designed camera movements and lighting profiles specifically tailored to harmonize with AI-generated visual elements.

    Sound Design

    Q Department crafted dynamic audio and sound design to ground the surreal, AI-rendered sequences in biological reality.

    “The film blends traditional live-action cinematography with AI-generated visuals to convey an internal, emotional medical drama.”