Every decade or so, the film industry gets handed a piece of technology and told it changes everything. Sometimes it does. More often, it changes some things and leaves the rest exactly where they were — usually the parts that matter most. We've been making things long enough to have lived through a few of these moments, and we want to share an honest read on where AI sits in that lineage, what it can actually do for filmmakers right now, and what we don't think it's going to do any time soon.
We've seen this movie before. Sort of.
When the dot-com wave hit, the prediction was that the internet would gut creative jobs. It didn't. It created entire categories of work that didn't exist a decade earlier — and yes, it also rearranged or retired some old ones. When the iPhone shipped a camera that fit in everyone's pocket, the prediction was that professional photographers were finished. They weren't. The bar for "good enough" rose, the pool of casual shooters exploded, and the people who actually understood light, lenses, and composition kept working — often with better gear, better clients, and a clearer reason to exist.
AI rhymes with both of those moments. The pattern is familiar: a new capability arrives, the loudest takes are catastrophic, and the actual shift is messier and more interesting. But AI is also a different beast. It doesn't just lower the cost of an old workflow the way digital cameras did; it generates content directly, at speed, in ways that pressure the cost structure of nearly every step of production at once. Pretending it's just "another camera" undersells it. Pretending it's the end of the craft overstates it. The truth is in the middle, and the middle is where the work is.
Short term: it's already in the kit.
The thing nobody on the AI hype circuit likes to admit is that the version of "AI in film" that's actually working right now is mostly unglamorous. It's transcription that turns a six-hour interview into searchable text in eleven minutes. It's auto-rotoscoping and noise reduction in DaVinci Resolve so a colorist can spend their afternoon on the shot that matters instead of the shot that's just in the way. It's storyboard sketches generated from a paragraph of script, useful not because they're final art but because they unlock a conversation faster. It's AI-assisted dubbing that compresses a localization pipeline by more than half. It's rough-cut suggestions that an editor accepts or — more often — overrides.
For a boutique studio like ours, this is meaningful. Tools that used to require a bigger crew or a bigger post budget are sitting on the same laptops we already own. That's a real win, and we'd be lying to say otherwise. AI is, in this practical sense, exactly what the optimists promised: another tool in the arsenal. We pick it up when it helps and put it down when it doesn't.
Long term: jobs will move, not vanish.
The honest long-term answer is: yes, some jobs are going to change shape, and a few will get squeezed out of the middle. The early casualties are tasks, not people — junior assistant work, dailies organization, basic VFX cleanup, simple voiceover localization, certain kinds of stock-image and concept-art commissioning. At the same time, new roles are already showing up on studio org charts: AI supervisors, virtual production technicians, prompt directors, likeness-rights managers, synthetic performance coordinators. If you squint, it's the same churn the industry has run through before, just compressed into a tighter timeline.
What doesn't go away is the part of the job that requires somebody in the room making decisions. Somebody has to know why this take is better than that one. Somebody has to feel the half-bar of laughter that lands before a downbeat and decide to leave it in. Somebody has to walk into a hotel ballroom in New Orleans, look at the light, and know which corner the camera belongs in. None of that is a prompt.
Earlier this year we shot Judith Owen's "That's Why I Love My Baby" — the lead single off Suit Yourself — live in the streets during Mardi Gras. Shooting a music video during Mardi Gras had been a long-held dream of Judith's, and we were honored to be the team she trusted to pull it off. The whole point of being out there was to put a real performer in the middle of a real city in full swing and catch what the city did back. Crowds shifted. Color spilled in from every direction. The light moved the way it always does down here — quickly, and on its own schedule. The shoot worked because a real crew read a real street in real time, and Judith met it head-on. You cannot prompt that. You cannot synthesize the specific gravity of a place you didn't plan, or the way a performance lands differently when it's actually inside the city that made it. An AI model can only imitate the look of New Orleans. Sure, it can describe a Mardi Gras day in detail, drawn from human experience — but human experience changes second by second. An AI model can never give us first-hand human experience of its own.
Hype vs. reality.
The hype says you'll generate a feature film from a paragraph by the end of next year. The reality is that generative video still struggles with character consistency across clips, still tops out at short durations, and still produces material that needs a human eye to be cut into something that means anything. Sora, the tool many people pointed to as the inflection point, was shut down earlier this year. Runway and its peers are doing serious work — but as tools that filmmakers steer, not tools that replace them.
The hype says AI will replace human creatives. We don't believe it will. The reality is that AI is excellent at producing competent, average, on-template content — and the entire premise of Fish Pot Studios is that competent, average, on-template content is not what people remember. The world still wants singular work. AI does not yet make singular work. It accelerates the parts of the process that surround it.
Will some jobs be replaced? Probably. Will AI end humanity? It's too early to say with certainty, but no — that's not the bet we'd take, and it's not the conversation that's useful for anyone trying to make a film this year.
The audience can tell.
The hype tends to miss the audience, and the audience is paying closer attention than the industry wants to admit. Gen Z in particular — a generation that grew up online and has been trained on a decade of feeds, ads, and algorithmic noise — has developed an almost reflexive ability to spot a synthetic face, a synthetic ad, a synthetic story. Recent research keeps landing on the same point: roughly four in ten Gen Z viewers feel negative about AI-generated advertising, a majority prefer human-made content when they can identify it, and the gap between what ad executives think Gen Z feels and what Gen Z actually feels keeps getting wider, not narrower.
The takeaway isn't that audiences are anti-AI. They're anti-inauthentic. They're pro-effort. They scroll past the thing that looks like it was assembled in twelve seconds because they've already seen a thousand of those, and they stop on the thing that feels like a person made it for them. Authentic human connection — the look on a face that wasn't art-directed, the cadence of a voice that wasn't smoothed — is doing more work in 2026 than any production trick on the market.
Apple's new Apple TV intro is a useful case study. The two-second sonic logo that now opens every Apple TV+ title was made by TBWA Media Arts Lab using actual cut and polished glass filmed against shifting fields of color — no CGI, no 3D modeling, no generative anything. Apple could have synthesized that ident for a fraction of the cost and called it a day. They didn't. Tor Myhren, Apple's VP of marketing communications, framed the choice in terms of "tactile detail and camera-centric aesthetics," which is a polite, corporate way of saying that even the most computational company on earth knows audiences can feel the difference between a thing that was made and a thing that was generated. Tellingly, even the AI companies see it: OpenAI shot its first major ChatGPT campaign on 35mm film, and Anthropic's recent Claude campaign was a cinematic short. When the companies selling AI are reaching for human craft to make their own commercials, that's the market telling you what it actually wants.
This matters for everyone making things for a screen. Whatever the production stack looks like, the work still has to feel touched. The studios and brands that figure this out — using AI for the parts of the process audiences never see, and keeping humans on the parts they do — are going to be the ones who earn attention worth keeping. The ones that lean entirely on the machine are going to keep wondering why their content gets scrolled past.
Where we land.
AI is a tool. A powerful one. A different one. We're using it where it helps, ignoring it where it doesn't, and watching the rest carefully. The studios that are going to come out of this decade in the strongest position aren't the ones that bet everything on the machine, and they aren't the ones that refuse to touch it. They're the ones who keep their hands on the work, use whatever sharpens the craft, and stay clear-eyed about which parts of filmmaking a machine can do and which parts still require a human standing in a room, making a call.
We're in the second camp. We'll let you know what we learn.