AI in Documentary Practice
A hands-on three-day workshop examining how to leverage AI technologies in documentary filmmaking. Moving beyond theory to focus on concrete applications - what works, what doesn't, and why - this intensive course explores the thoughtful integration of AI into documentary practice, from research and development through production and post-production.
While AI technologies offer remarkable new creative possibilities and production workflows, their implementation requires careful consideration of both artistic and ethical implications. Through hands-on exercises and critical discussions, participants will explore current AI capabilities while maintaining creative vision and documentary values. This workshop provides a grounded, practical perspective that will help filmmakers and producers evaluate and integrate these technologies effectively in their own projects.
Across three intensive days, participants move from the foundations of AI and computational film analysis (Day 1), through multimodal models and image generation for documentary (Day 2), to video generation, voice synthesis and integrated post-production workflows (Day 3). Every block is anchored in a hands-on exercise on the participant's own material — by the end of the workshop, each filmmaker leaves with a 60-second pitch reel produced during the course and a personal roadmap for integrating AI into their next project.
Program
Day 1 - Wednesday 16 September 2026
Research & Development with AI
Morning Session (09:00 – 12:00) - Block 1: Foundations & Analytical AI
Introductions and calibration with participants' own projects
A short history of creativity, automation and AI
What AI is - myths, realities, and the cognitive-load research
Training models - the invisible labour
Discriminative AI in documentary practice through case studies
Computational film analysis: shot detection, embeddings, semantic search across rushes
Hands-on: reading footage at scale on a sample documentary sequence
Lunch break (12:00 – 13:00)
Afternoon Session (13:00 – 17:00) - Block 2: LLMs, Prompting & RAG
LLMs (Gemini, GPT, Claude) - origins, architecture, model variables
Anatomy of a prompt - the six blocks (role, context, task, format, constraints, guardrails)
Cloud vs. local models (LM Studio + Gemma) - privacy and cost trade-offs
Retrieval-Augmented Generation (RAG) and document analysis with NotebookLM
Hands-on: build your own research assistant over a corpus of your own documents
Day 2 - Thursday 17 September 2026
Production Enhancement & Image Generation
Morning Session (09:00 – 12:00) -Block 3: Multimodal AI & Image Fundamentals
Multimodal Large Language Models (MLLMs) -what changes when models can see
Working with AI for visual content: indexing, captioning, archive review
Three model levels (chat, reasoners, agents)
Image generation in 2026 -comparative review of leading models (Flux, GPT Image, Nano Banana, Seedream)
Hands-on: prompt engineering for visual creation
Lunch break (12:00 – 13:00)
Afternoon Session (13:00 – 17:00) - Block 4: Advanced Image Techniques for Documentary
Visual style adaptation for documentary aesthetics
Image upscaling and enhancement of archival material
Continuity, character reference and visual coherence across a sequence
Hands-on: producing a coherent set of nine documentary-style images for a sequence in your own project
Day 3 - Friday 18 September 2026 (09:00 -17:00)
Video & Audio Integration
Morning Session (09:00 – 12:00) - Block 5: Video Generation for Documentary
Video generation fundamentals (text-to-video, image-to-video, video-to-video)
AI Mapper - Artefacto's framework for the visual categories of AI cinema
Selective animation of static archival material with Google Flow
A four-phase generative video pipeline: character base → reference grid → frame breakdown → shot-by-shot animation → edit
Hands-on: animate the nine-frame sequence produced on Day 2 into a 15-second piece
Lunch break (12:00 – 13:00)
Afternoon Session (13:00 – 17:00) - Block 6: Audio Integration & Closing
Voice synthesis and voice cloning ( Instant Voice Clone)
Audio enhancement of damaged archival recordings
Talking avatars and synthetic presenters - when they help, when they break the documentary contract
Performance transfer and motion control (Motion Control / Match Video)
Provenance and transparency: C2PA Content Credentials and SynthID watermarking
Hands-on: produce a 60-second documentary pitch reel combining generated image, video and synthetic voice
Final discussion: integrating AI in documentary workflows · personal roadmaps · ethical and technical limits
What will you learn?
- State-of-the-art map of AI tools for documentaries - what's real, what's hype, and what's worth integrating in 2026.
- Advanced research techniques with LLMs and Retrieval-Augmented Generation over your own archives, transcripts and bibliography.
- Computational film analysis - using vision models to read footage at scale (shot detection, character tracking, semantic search).
- Production workflow optimisation - image generation, multimodal models, and visual style adaptation tailored to documentary.
- Post-production with AI - video generation, voice cloning and audio enhancement, integrated without losing documentary integrity.
- An informed view on ethics, provenance and future trends - C2PA Content Credentials, SynthID, copyright, and how to talk to commissioners about AI use.
Who should attend?
This intensive workshop is designed for documentary filmmakers and producers who:
Want to explore AI's potential in documentary filmmaking from a practitioner's perspective.
Seek to enhance research, development and production efficiency through AI.
Are interested in cutting-edge tools currently shaping the industry.
Want to stay competitive in a fast-changing landscape — and make informed decisions about when not to use AI.
The workshop will be conducted in English and focuses on practical, hands-on experience. Participants must bring their laptops.
Application process
To apply, please submit a motivation letter and your CV to efteruddannelsen@filmskolen.dk
Workshop participants who work professionally in the Danish film industry can apply for financial funding from the Danish Film Institute's pool for continuous training and competence development. The funding is for those who work or have worked professionally in the Danish film industry and the people who have a relevant education in film production.
The institute funds course fees up to DKK 25,000.