[ Standard / Evaluation ]
The 51% AI Participation Standard
Eligible AI-native films show AI as the leading creative force across weighted dimensions of the production—not a lone VFX statistic or incidental assist.
What 51%+ AI participation means
The threshold is an expert review of holistic creative dominance, not a spreadsheet cell or single render pass percentage.
- ▸Jurors synthesize documentary evidence—including creative journals, timelines, prompts, selects, mix stems, VFX turnovers, and graded masters—to judge where AI materially led each dimension.
- ▸The number refers to dominance across storytelling dimensions collectively. A film does not qualify by maxing one category while ignoring the rest—it must convincingly surpass the bar in aggregate.
- ▸Human authorship remains vital: directors and producers steer ethics, taste, and final approval. The standard simply clarifies where AI—not traditional departments alone—drives inventive outcomes.
Six creative dimensions evaluated
Each dimension earns a juried weight toward the whole. Together they sum to 100%.
Total weight100%
Writing (screenplay)20%
- ·Story architecture shaped with AI-assisted ideation.
- ·Character arcs refined through iterative AI drafts.
- ·Dialogue experimentation beyond light proofreading.
Direction20%
- ·Shot sequencing planned with AI visual references.
- ·Performance tone explored via generative rehearsals.
- ·Narrative blocking informed by diffusion or sim tools.
Generation (image / video)20%
- ·Principal imagery created or heavily rebuilt with generative pipelines.
- ·Look-dev cycles driven by diffusion / video models.
- ·Synthetic environments that define on-screen reality.
Editing15%
- ·Structural iterations generated or ranked with AI assistance.
- ·Temp scores or rhythm studies created synthetically.
- ·Alternate cuts explored at scale before picture lock.
Sound15%
- ·Stem-level design co-created with generative audio tools.
- ·Adaptive score palettes shaped by AI timbre explorers.
- ·Soundscapes impossible to source solely from library foley.
Post (VFX · finishing)10%
- ·Compositor-grade passes leaning on neural rendering.
- ·Color decisions iterated through AI-assisted looks.
- ·Mastering tweaks directly tied to generative rework.
How to self-assess your film
- 01Would removing AI workflows materially change the story beats or emotional arc?
- 02Did diffusion, video, or neural tools define the iconic visuals audiences remember?
- 03Were editing timelines or pacing explored through AI-generated alternates?
- 04Does the soundtrack hinge on synthetic instruments, generative ambience, or AI-assisted cues?
- 05Did VFX finishing or grading rely on generative rework—not just cleanup?
- 06Can you assemble dated evidence (journals, exports, renders) proving AI leadership per dimension?
What HKAIIFF will verify
- ■Signed creative diary tying major decisions to AI explorations versus traditional plates.
- ■Tool logs / render stamps / cloud job IDs showing model usage across departments.
- ■Side-by-side evidence of AI output ratios blended into the locked master—for example stem splits or VFX turnovers.