Photography & Generative AI
A clear-eyed hybrid workflow
My photography expertise is now augmented by generative AI tools. I design complete campaigns around a specific subject — emotional, sensory, or technical — combining real photography with assisted generation. Not a gimmick: every image answers a goal we set together. The anchor stays the same — moody cinematographic imagery, human expression, realism.
Where AI works
- Moodboards and visual exploration — two or three directions visualized in a few hours.
- Scenarios that don't exist yet — an R&D process, a hall not yet built, a prototype to be staged.
- Lighting tests — simulate three moods before mobilizing a team and a drone.
- Extensions and corrections — extend a background, clean up a scene, create variations from a real photo.
Where AI breaks
Your real products, machines and team members. As soon as a detail matters — a logo, a face, the exact geometry of a part — AI hallucinates. For a moodboard, that's fine. For a tech sheet or a website where your customers recognize your facilities, it's a deal-breaker. That's why I position AI as a complement, never a replacement.
The hybrid workflow in practice
- Brief & exploration in AI — Midjourney or Stable Diffusion to frame the mood.
- Real shoot — on site, with your team, your machines, your process.
- Assisted post-production — Photoshop with Generative Fill and Magnific for extensions and finishing.
Tools I use
Midjourney, Stable Diffusion, Magnific (upscaling and finishing), Photoshop with Generative Fill, ComfyUI for specific workflows.