Martin Scorsese joins Black Forest Labs as advisor

💡High-profile industry adoption of AI tools is fueling a major debate on the future of creative labor in film.
⚡ 30-Second TL;DR
What Changed
Martin Scorsese appointed as partner and advisor to Black Forest Labs in 2025.
Why It Matters
This partnership highlights the growing tension between traditional creative workflows and AI integration in Hollywood. It signals a shift where high-profile directors are actively shaping the development of generative tools, potentially accelerating AI adoption in film pre-production.
What To Do Next
Evaluate Black Forest Labs' API capabilities for your own creative workflows to see if their image generation quality meets professional production standards.
Key Points
- •Martin Scorsese appointed as partner and advisor to Black Forest Labs in 2025.
- •Scorsese utilizes AI-generated storyboards to communicate his creative vision to production crews.
- •The move has triggered significant backlash from film industry professionals concerned about artistic labor.
- •Black Forest Labs specializes in text-to-image generative AI technology.
🧠 Deep Insight
Web-grounded analysis with 20 cited sources.
🔑 Enhanced Key Takeaways
- •Martin Scorsese's justification for using AI storyboarding centers on improved communication of his vision to the crew, potential time and cost savings, and reduced crew burden, drawing parallels to his past adoption of 3D for 'Hugo' and de-aging for 'The Irishman'.
- •The announcement of Scorsese's advisory role triggered specific backlash from film industry professionals, with artists like Karla Ortiz and Samuel Deats accusing him of undermining artistic labor and utilizing AI trained on potentially 'stolen work'.
- •Black Forest Labs, founded in 2024 by former Stability AI engineers responsible for Stable Diffusion, offers the FLUX.2 image generation model, which Scorsese specifically uses for pre-visualization.
- •Black Forest Labs is valued at $3.25 billion and lists Adobe, Mistral AI, and Nvidia among its clients, indicating a significant industry presence beyond just filmmaking applications.
📊 Competitor Analysis▸ Show
| Company/Product | Focus/Key Feature | Black Forest Labs (FLUX.2) |
|---|---|---|
| Image Generation | ||
| Midjourney | Artistic, visually striking images with distinct style. | State-of-the-art quality, speed, and controllability; strong prompt following, visual quality, image detail, and output diversity; production-grade consistency for characters and styles across assets. |
| Stability AI (Stable Diffusion) | Open-source, detailed, realistic images from text prompts; high customization. | Outperforms Stable Diffusion 3 Ultra in internal qualitative tests for visual quality and prompt following. |
| DALL·E 3 | High-quality image generation. | FLUX.1 models outperformed DALL·E 3 HD in internal qualitative tests. |
| Filmmaking Pre-production | ||
| Shai Creative | All-in-one script-to-storyboard workflow, AI video generator for animatics, shotlist creator. | Specializes in generative image technology for storyboarding and pre-visualization, with a focus on director's vision communication. |
| Runway Studios | AI video generation, full production workflow, AI Film Festival. | Focuses on high-quality, consistent image generation for pre-production, specifically storyboarding. |
| Adobe Firefly | AI-generated storyboards, generative fill, changing video clip attributes. | Offers production-grade AI image generation and editing model (FLUX.2) with features like character consistency and high resolution. |
🛠️ Technical Deep Dive
- Black Forest Labs' FLUX models (FLUX.1, FLUX.2) utilize a 'hybrid architecture' that combines transformer and diffusion techniques.
- The models are trained using 'flow matching,' a general method for generative models that encompasses diffusion.
- FLUX.1 [dev] and FLUX.1 schnell are 12 billion-parameter models.
- Performance and hardware efficiency are enhanced through the incorporation of rotary positional embeddings and parallel attention layers.
- FLUX.2 [klein] models, available in 4 billion and 9 billion parameters, are optimized for speed, capable of generating images in under a second on modern hardware like Nvidia GB200, and can fit within approximately 13GB of VRAM on consumer GPUs (e.g., RTX 3090 or 4070).
- The FLUX.2 models support output resolutions up to 4 megapixels and any aspect ratio, maintaining reliability from low-resolution drafts (400px2).
- They offer 'production-grade consistency' for characters and styles across numerous assets and can reference up to 10 images to maintain character consistency.
- FLUX.1 Kontext is a suite of models enabling in-context image generation and editing by allowing users to prompt with both text and images.
🔮 Future ImplicationsAI analysis grounded in cited sources
⏳ Timeline
📎 Sources (20)
Factual claims are grounded in the sources below. Forward-looking analysis is AI-generated interpretation.
Weekly AI Recap
Read this week's curated digest of top AI events →
👉Related Updates
AI-curated news aggregator. All content rights belong to original publishers.
Original source: The Guardian Technology ↗

