Brix36Visit Website →
← Back to blog

AI Short Films: Watch Festival Picks, Then Make Your Own

October 17, 2026
AI Short Films: Watch Festival Picks, Then Make Your Own

An AI short film is a short narrative or experimental film built with generative AI tools, and you can watch it, browse curated collections of it, or make your own. We point you to festival hubs such as the Astana AI Film Festival, explain what the U.S. Copyright Office expects from AI-assisted work, and show how Brix36 fits if you want to produce.

TL;DR:

  • Reuse the same character and location references, lock aspect ratio and frame rate early, and generate audio first when visual timing depends on it.
  • Budget for repeated generations, since costs usually follow credits or outputs and retakes can consume more than the first render.
  • The U.S. Copyright Office allows registration of the overall film even when generated elements are not protected, but applicants must disclose AI components.
  • Get written consent before cloning a voice or likeness, avoid unlicensed inputs, and seek legal review when footage resembles a real person or protected work.
  • Before submitting, check each festival’s rules for runtime, language, subtitles, eligibility, and AI disclosure because requirements differ by program.
Brix36
Turn Your AI Film Idea Into Reality
Brix36 helps creators make films with AI tools and one-tap looks, simplifying production for artists and filmmakers at any skill level.
Explore Brix36

Table of Contents

Where can you watch AI short films right now?

Think of the viewing landscape as three kinds of hubs, each with a different payoff.

  • Festival channels: Many festivals post selected films to their own YouTube channels or program pages, so you see work a jury already filtered.
  • Curated sites and aggregators: These collect shorts under searchable tags, which suits browsing by style, runtime or theme.
  • Platform playlists: Creators and communities build playlists around AI-native work, usually the fastest way to find new premieres.

Festival hubs tend to feature a tight selection with credits and tool notes. Aggregators lean on volume and tags. Playlists reward whoever curates them well, so follow a few curators rather than one giant list.

For search, combine "AI short film" with a festival name, a year, or a style word such as "animated" or "experimental." Hashtags like #aifilm and #generativefilm surface AI-native shorts on social platforms, and the phrase "made with generative AI" often appears in video descriptions.

What should you notice in recent AI short films?

Watching an AI short with a notebook teaches more than any tutorial. Pick three to five recent films from the hubs above and look for the choices below.

  • Full-AI direction: Films where every shot is generated show how a director guides a model through prompts, references and selection rather than through a camera.
  • Stylized palettes: Strong color rules hide the visual drift that generated footage tends to have, so a limited palette is often a craft decision, not a limitation.
  • Hybrid authorship: Some shorts blend human performance, hand-drawn art or live footage with generated elements, and the seams reveal the workflow.
  • Short runtimes: Many of the strongest entries run only a few minutes, because tight scenes keep characters and settings consistent.

Together, these point to a new cinematic grammar: shots built from references and intent instead of coverage, and edits that hide the model's weak spots while showcasing its strengths.

Pro Tip: Watch each short twice: once for story, then with the sound off to study shot length, motion and color. Write down which shots feel generated and which feel designed, then copy the designed ones in your own project.

How do you make an AI short film, step by step?

The workflow below is tool-agnostic, so it works whichever generators you choose. Our AI video workflow guide covers how to connect several tools without rebuilding your pipeline each time.

  1. Concept and script: Write a one-page story with a single character goal and a handful of locations. Scripts for AI work best when every scene can be described in a sentence.
  2. Reference assets: Gather or generate character sheets, location boards and a palette. Text-to-image models are the usual tool here, and these references anchor everything that follows.
  3. Generation: Create frames and motion using image-to-video or motion models, and voices with speech synthesis. Generate several takes per shot and keep the best.
  4. Assemble and edit: Cut in a standard editor, then use AI-assisted tools for cleanup, upscaling, background removal or compositing. If you need a clean cutout, see how to remove a video background without a green screen.
  5. Finalize and export: Color match across shots, mix audio, add credits with an AI disclosure line, and export for each destination.

Consistency is the hard part. Reuse the same reference images and descriptive phrases for every shot of a character, lock your aspect ratio and frame rate early, and generate audio first when timing matters so visuals can follow it.

Pro Tip: Build a "shot bible" document with the exact wording that produced your best character and location frames. Pasting those lines into every new prompt does more for consistency than any setting.

Which festivals and communities showcase AI-native films?

A growing ecosystem now treats AI work as its own category. These are good places to watch, learn and eventually submit.

  • Astana AI Film Festival (AAIFF 2026): Runs competitions and curated selections, and publishes submission windows and eligibility requirements.
  • WAiFF 2026: Frames generative AI as an emerging artistic language and programs short and hybrid formats across many categories.
  • Marbella International Film Festival: An established festival that lists a dedicated slate of AI films, each with runtime and country details.
  • Shaike Film Festival: Highlights collaboration between traditional cinema practitioners and AI-native creators.

Signals that a festival takes AI seriously include a dedicated category, a named jury, and programs that mix formats. To follow premieres and discussion, subscribe to festival feeds and YouTube channels, then watch which creators recur across programs.

What does a first AI short film project look like?

A realistic first project fits into a few weeks and a small budget. Use this plan and adjust the timings to your schedule.

  1. Prototype (days 1–3): Write the one-page script and generate a handful of test frames.
  2. Reference lock: Finalize character, location and palette references.
  3. First render: Generate every shot, with several takes each.
  4. Edit: Assemble, trim and add temporary sound.
  5. Polish: Color, mix, titles and credits.
  6. Release: Export, write the metadata and submit or post.

Costs usually follow a credit or per-output model, and most of the budget goes to regeneration: the more takes you need per shot, the more you spend. Plan for retakes rather than one clean pass. If you want a tour of how credits work, our credits page explains the pay-as-you-go model.

For a festival or demo, prepare these minimum deliverables:

  • A finished short with a clear runtime
  • A credits line that discloses which parts used AI
  • A title, logline and basic metadata
  • A trailer or still frames for submission forms

What rights and responsible-use rules apply?

Copyright offices are actively examining AI outputs. The U.S. Copyright Office has published multi-part analysis covering copyrightability of AI-generated outputs, training data and digital replicas. When registering a film that contains generated material, applicants must disclose the AI components, while the registration of the overall film can stand even if generated elements themselves are not protected.

The Copyright Office has documented a public inquiry into copyrightability of AI-generated outputs and issues around training data and digital replicas, so treat your AI components as something you may need to explain.

Production guidance from Netflix's partner help center recommends getting consent for likeness or voice, avoiding final deliverables that reproduce protected works without approval, and escalating for legal review when outputs involve talent likeness or copyrighted material. A practical checklist:

  • Disclose AI usage in your credits
  • Avoid inputs you are not licensed to use
  • Get written consent before cloning any voice or likeness
  • Ask a lawyer when a shot resembles a real person or protected work

How do you tell a better story when a model makes the shots?

Write for what generation does well. Models handle mood, texture and single striking moments far better than long continuous action, so build stories from images that carry meaning on their own: a face at a window, an empty room after an argument, a city that feels slightly wrong.

Keep the cast small. One or two recurring characters are far easier to hold steady than a crowd, and a narrator or voice-over can carry exposition so you do not need to show everything. Use limited locations and repeat them from new angles.

Embrace the dreamlike quality instead of fighting it. Stories about memory, loss, transformation and surreal worlds make visual drift feel intentional, while stories that demand strict realism expose every flaw. Pace matters too: cut early on shots that start to wobble, and let the sound design do the heavy lifting where visuals are thin.

Finally, choose a rule and obey it. A single color scheme, a repeated visual motif or a fixed camera distance gives viewers a pattern to trust, and that sense of design is what separates a film from a pile of generated clips.

How do you blend AI-generated elements with live-action footage?

Hybrid work starts on set, not in software. Shoot clean plates of empty backgrounds, record lighting notes and keep the camera locked off or tracked, so generated elements have something stable to sit on.

Match the basics first: perspective, light direction, color temperature and grain. A generated character dropped into a plate looks fake mostly because the shadows and noise do not match, not because the model failed. Add film grain, slight blur and matched color grading across both layers to glue them together.

Isolate elements cleanly. Rotoscoping and background removal let you separate people or objects and place them in generated environments, and our guide on removing a video background without a green screen walks through different approaches. Run generated footage through the same finishing steps as your live footage: stabilization, denoise, then a shared grade.

Use AI where the camera is weakest. Generated skies, set extensions, background crowds and impossible spaces cost far less than building them, while close performances are usually best left real. Keep generated shots short and cut away before inconsistencies show.

Which AI models and algorithms power short film production?

You do not need the math to work with them, but knowing the categories helps you pick tools and diagnose problems.

Diffusion models power most text-to-image and many video generators. They start from noise and refine it toward an image that matches your prompt, which is why they respond well to detailed descriptions and reference images.

Video and motion models extend those ideas across time, producing short clips from text or from a still image. They are strongest on brief shots, so expect to stitch many clips together.

Speech and voice synthesis converts text into spoken audio and can match a reference voice, with consent. Lip-sync models then drive a face to match the audio, which is central to music videos, as our guide to making an AI music video from your song shows.

Language models help with scripts, shot lists and prompt writing, acting as a fast brainstorming partner.

Assisted editing tools handle upscaling, frame interpolation, denoising, rotoscoping and compositing, and often matter more to the final look than the generator itself.

Think of these as a toolbox rather than a single product: most finished films combine several categories.

Five AI tool categories used in short film production

How do you optimize AI outputs for social, festivals and streaming?

Each destination rewards a different master file, so plan exports before you generate.

Social media favors vertical or square framing, a hook in the first seconds and captions, since many viewers watch muted. Generate or crop in the target aspect ratio from the start rather than reframing later, which can cut off faces.

Festivals usually want a full-length master with clean audio, stable resolution and clear credits. Check each program's rules on runtime, language, subtitles and AI disclosure before you submit, and keep your disclosure wording consistent. The Astana AI Film Festival publishes submission windows and eligibility requirements, which is the kind of detail to confirm early.

Streaming and partner platforms tend to expect high-quality deliverables, strong audio mixes and documentation about how generative tools were used. Follow their guidance on consent and legal review.

Across all three, upscale and denoise generated footage, normalize audio levels, and watch for artifacts such as flickering textures or warped hands at full size. Export a high-quality master first, then derive smaller versions from it so you only fix problems once.

How do you optimize AI outputs for social, festivals and streaming? — overview diagram

How should you collaborate with AI tools as a human creator?

Treat the model as a fast, tireless assistant that never has taste. You supply the intent, the selection and the rejection; it supplies options for animators and VFX artists. The creative act lives in choosing among generated takes and refusing the ones that are merely impressive.

Set roles clearly if you work with a team. One person owns the visual references, another owns story and edit, and someone is responsible for rights and disclosure. Shared documents with prompts, seeds and approved references keep everyone aligned and make results repeatable.

Work in short loops: generate, review, adjust, repeat. Save versions so you can return to a take that worked, and keep notes on what changed between them. Review with fresh eyes or another person, because it is easy to fall for a shot that looks great in isolation but does not serve the scene.

Keep humans in charge of performance, voice and emotional beats whenever you can, and be honest about where AI contributed. Clear credits help explain your process to festivals, platforms and audiences, and connecting with other creators is a way to learn what works.

Why we believe AI short films deserve your experiments

I see AI as a creative partner, not a replacement for taste. The most interesting thing about these films is that a new grammar is emerging: shots built from references, edits that hide a model's limits, stories shaped around what generation does beautifully. We built Brix36 to make professional production accessible through one-tap looks and a credits model. Make something small, share it, and learn in public.

— simeon

Make your AI short film with Brix36

Ready to produce? Our short film maker is a pay-as-you-go option for filmmakers, freelancers and small studios who want to finish a short quickly, without a subscription.

Brix36

  • AI video tools for building scenes and sequences
  • Artist DNA to keep one artist's look consistent across projects
  • One-tap looks for professional styling without technical skills
  • Credits you buy and spend only when you create, with details on our pricing page

Start with a one-page idea and spend your first credits on a single scene.

FAQ

Can I create a short film with AI?

Yes. You can write a script, generate frames and motion, add synthetic voices and edit the result like any other film. Start small, with one character and a few locations, and disclose which parts used AI.

Where can I watch AI short films?

Festival programs such as the Astana AI Film Festival and the Marbella International Film Festival's AI selection are good starting points, along with festival YouTube channels and curated playlists. Searching a festival name plus "AI short film" surfaces most of them.

The legal position can vary by country and is still developing. The U.S. Copyright Office is actively examining AI outputs, so disclose AI components, get consent for any likeness or voice, and seek legal advice for borderline cases.

Do festivals accept AI-generated films?

Many do, and some are built around them. Programs like WAiFF 2026 feature short and hybrid formats, while others add AI categories. Always check each festival's eligibility and disclosure rules before submitting.

Sources