Riccardo Fusetti first hit our pages with his experimental short Generation – an innovative exploratory AI-generated story of human existence – and we’re delighted to welcome him back with his latest piece, Editorial. Working with the vibrant British-Bulgarian artist/actor Victoria Valcheva, Fusetti has crafted a visually dense, rapid fire short which moves from a frivolous, buoyant fashion editorial into something else entirely. The project emerged as an exploration of contrasting extremes of femininity, blending lo-fi aesthetics with high-end production values and Valcheva’s self-shot footage to create a disorienting yet cohesive narrative experience while adopting a visual language from a wide mix of formats, including AI-generated imagery. This technical diversity isn’t merely stylistic but thematically essential, and Fusetti’s approach allowed for the construction of a dense, multifaceted world far beyond the limitations of the project’s limited budget. Joining us for Editorial’s premiere, Fusetti tells DN how he leant into the extreme dichotomies of womanhood, the added value incorporating genuine footage from Valcheva’s personal phone and camera archives brought to the film and his methods for contending with the often inconsistent results produced by AI tools.
Did you begin this project with a definite idea of where you wanted the narrative to go?
Absolutely, I really wanted the audience to feel like they were watching a fun, full-fledged, carefree and stylish day in the life fashion editorial and then subvert it and take it to a dark and surreal place. I intentionally let the first part linger to make the dark turn hit even harder. The film was born out of a need to create something fun and uncompromising – something that felt grand in scale despite being made on a shoestring budget. From the start, I knew I wanted to capture micro-glimpses of a character’s life, racing through quick cuts at incredible speed, yet forming a cohesive narrative when assembled.
It was essential to showcase the full spectrum of our character and her femininity, I wanted the film to be about juxtaposing extremes: the lo-fi with the high-end, the real with the fictitious, we see our model, wonderfully portrayed by Victoria Valcheva, as stunning, cute, and fun, yet also incredibly dark and horrifying. The film also doesn’t take itself too seriously and it’s supposed to be a fun and wild rollercoaster.
The descent into madness came quite abruptly but the depth we go down into gets gnarly – did you just want to see where it would take you?
Indeed. At first, I thought we would lean more into the surreal and elevated elements, but as I kept experimenting, I found myself loving the idea of pushing things further into the gnarly. It was a lot of fun, and I think it makes our character much more interesting and layered. Victoria and I really wanted to push the character to both extremes – cutesy and horrific.
As I kept experimenting, I found myself loving the idea of pushing things further into the gnarly.
How did you go about integrating your lo-fi and the more high-end parts of Editorial?
This approach was reflected in the way we shot the film, using a mix of high-end 16mm film, Bolex, Arri Alexa, DSLRs, MiniDV, phone footage, and AI-generated footage. To expand the film’s scope and create a rich, immersive narrative world, we incorporated authentic footage from Victoria’s personal archive – small clips from her phone and camera. These clips are truly genuine, captured in the past with no intention of being used in a film. I felt they were the perfect foundation for building our own character. I also thought it would be exciting to juxtapose these completely real and unfiltered moments with fully ‘fake’ and entirely AI-generated clips.
What truly ties everything together is the sound department. Alexander Wells did an incredible job composing all the music and crafting the visceral sounds that immerse you in the world. Francis Qureshi at Wild Island Films then masterfully graded the footage, somehow making it all feel cohesive and belonging to the same world.
Most Popular
Victoria and I started developing and discovering the character together, and it was so much fun to improvise and explore. We would meet and shoot quick, lo-fi moments on a phone, DSLR, or Bolex. I also asked her to continuously film herself in character throughout the months of production, building an archive of clips for the edit. This spontaneous approach was paired with one and a half shoot days where things were planned out, storyboarded and shot on high-end gear to fill in the gaps. Our DoP Toby Goodger did an incredible job on these more composed and thought out shots. We were incredibly lucky that Kodak gifted Victoria with some film stock that she donated to the project. It was very exciting to shoot 16mm on such a small project and I am incredibly happy with how it looks. The shooting process was incredibly playful. Even during the more structured shoots, we were just experimenting and having a great time on set. Some ideas were downright silly and goofy, and many shots didn’t make the final cut. Overall, I ended up with almost 10 hours of footage in total.
I also asked her to continuously film herself in character throughout the months of production, building an archive of clips for the edit.
With AI-generated content being such a controversial topic in filmmaking, what considerations guided your implementation of the tools, and how did you navigate the balance between artistic experimentation and authenticity to your concept?
It was crucial to have the film play out like a very traditional fashion and lifestyle editorial at first but I knew straight away that it would have to descend into pure surreal madness and horror towards the end. This was achieved by generating and photoshopping some stills of Victoria and then using some AI image-to-video models to animate said stills. I ended up with almost 80 AI VFX shots sprinkled throughout the film. I was experimenting with new AI tools and wanted to try my hand at achieving photorealistic results. I know it’s a divisive topic, and I have my own doubts about these tools, but I saw this as an opportunity to test them out. In the end, I’m very happy with the results – I feel I was able to truly expand the world and give it a greater sense of scale.
While the tools are improving, they’re still far from perfect. This entire piece was carefully designed to work within the significant limitations of AI-generated videos, and it took months of trial and error to reach this stage. The very nature of the short complements the generated shots, each clip is so fast-paced that you never have time to focus on the imperfections. I used the AI shots as weird and extreme accents to elevate the film and take it somewhere it couldn’t have gone on our budget. The goal was always to make a very human portrait and keep it largely human-driven. Every shot is extremely intentional and it took countless hours and a lot of manual labour to get the AI VFX shots to work and I really don’t feel like we lost ‘authenticity’ with this workflow. Again, juxtaposing the ‘real’ and the completely synthetic was an important element of the film – the film is very self-aware and acknowledges this nature and the fact that we are using all sorts of tricks. There is a (hopefully funny) shot where we have Victoria in a greenscreen suit, glitching out and breaking the image – winking at the use of VFX even if we never really use a green screen.
I used the AI shots as weird and extreme accents to elevate the film and take it somewhere it couldn’t have gone on our budget.
What were the most significant technical constraints you encountered with the current generation of AI video tools, and what specific techniques did you develop to overcome or creatively incorporate these limitations?
Well, results were incredibly inconsistent, and most of the generated clips were completely unusable. It was a constant battle of tweaking prompts and rerolling until I got something usable. The general rule was to keep things simple and somewhat grounded – simple motion, dark environments. In the end, it’s all just sleight of hand and the edit does most of the work in terms of selling the shots.
What were the advantages of adopting so many different shooting formats and did you encounter any specific technical challenges when integrating these various formats during post?
I actually think that mixing all these formats is the key ingredient that makes the film work. I was trying to find a way to make our world feel big and alive on a no-budget production, and I knew Victoria’s personal archive would be the key to achieving that. I wanted a fast-paced, messy edit, and the multiple formats really helped add character and texture to the film.
Going back to the sleight of hand – the constant switching between high-end and very lo-fi footage was also a way to blend in the low-resolution AI-generated clips. If the entire short had been shot with high-end gear, those clips would have stood out, but by mixing them with all kinds of footage, they seamlessly become part of the world.
How has your understanding and relationship with AI changed since your 2022 piece Generation?
Things have changed quite a bit since then, but I wouldn’t say my opinion has dramatically shifted. I find the tools fun and full of potential, but I see them as something to be used alongside existing workflows, not as a replacement. To be honest, I find the discourse around AI exhausting – there’s so much hype that feels excessive and divisive. I consider myself a filmmaker – traditional filmmaking is what I do, and it’s where the most fun is. I like using the right tool for the right purpose, but I don’t see traditional filmmaking disappearing anytime soon.
What are you looking to explore next?
I have a couple of projects in early development that I would love to get done this year and I’d like to push more towards longer narrative form.
