The annual AI Film Festival organized by Runway has transformed from a "crazy idea" to a significant showcase of artificial intelligence in filmmaking, with submissions skyrocketing from 300 in its 2023 debut to approximately 6,000 entries in 2025, reflecting the explosive growth of AI-generated video creation across the creative landscape.
Runway's Gen-3 Alpha represents a significant leap forward in AI video generation technology, serving as the foundation for the creative tools showcased at the AI Film Festival. This next-generation foundation model excels at producing videos with exceptional fidelity, temporal consistency, and realistic human characters capable of complex movements and emotions.12 Trained on both videos and images, Gen-3 Alpha powers Runway's suite of creative tools including Text to Video, Image to Video, and advanced control features like Motion Brush and Director Mode.1
Key innovations include fine-grained temporal control through highly descriptive captions that enable precise keyframing and imaginative transitions,2 superior consistency that prevents morphing between frames,2 and the ability to generate expressive human characters with natural movements.13 The model also supports longer video durations of up to 10 seconds (compared to Gen-2's 4-second limit) and introduces new capabilities like Act One Animator for character animation and Video-to-Video generation for transforming existing footage into different styles.4 These advancements have dramatically expanded the creative possibilities for filmmakers participating in the festival.
Jacob Adler's "Total Pixel Space" emerged as the standout winner at the 2025 AI Film Festival, capturing top honors with its hypnotic 9-minute exploration of digital imagery mathematics.12 The film poses a fascinating question: how many possible digital images can exist? Through elegant mathematical reasoning, Adler demonstrates that every conceivable digital image exists as coordinates in an abstract space, calculating that the total number of possible images is "a seven followed by over 7.5 million digits" – a number vastly exceeding the estimated atoms in the universe.3
The film's mesmerizing visuals traverse from recognizable scenes to abstract patterns, illustrating how "total pixel space contains every possible digital image, including films of your entire life, every life you never lived, and the lives of every creature or object that ever and never existed."3 Released in December 2024, the short film gained significant attention before its festival triumph, accumulating thousands of views across platforms and establishing Adler as an innovative voice at the intersection of mathematics, philosophy, and AI-assisted filmmaking.45
Mixed media animation combines multiple visual techniques to create dynamic, textured content that stands out in today's digital landscape. This approach merges live-action footage, 2D and 3D animations, text overlays, and special effects into cohesive visual narratives1. For filmmakers experimenting with AI tools, mixed media offers unique opportunities to blend the digital and physical worlds, creating works with both technological innovation and handcrafted authenticity.
The process often involves breaking videos into individual frames for analog editing with paints, crayons, stickers, fabrics, or other physical materials2. These manually enhanced frames are then scanned and reassembled into animations that preserve the texture and tactile qualities of physical media while incorporating AI-generated elements23. This hybrid approach allows creators to balance AI's capabilities with human artistic expression, resulting in visually rich productions that combine the precision of digital tools with the organic imperfections and creative spontaneity of traditional art techniques34.