AI on the Lot 2025: What creative AI adoption looks like up close

The KoobrikLabs team spent a few days at AI on the Lot last week – a gathering of filmmakers, technologists, executives, and innovators all trying to figure out, in real time, what AI means for the future of storytelling.

It’s the kind of event where you might watch a keynote about GPUs optimized for generative workflows, then sit in on a lab where someone’s prompting a model to light a dragon fight scene with cinematic accuracy. It’s a lot. In a good way. Mostly.

These types of conferences are always a balancing act. You get a front-row seat to what’s coming; new tools, new workflows, new roles (like “prompt artist,” which came up during the Amazon MGM Studios panel). But the speed of innovation can feel like it’s outpacing the frameworks (creative, legal, or organizational) that are supposed to support it.

There were sessions where I was fully locked in. Doug Shapiro’s talk on the critical role of infrastructure in enabling AI adoption was a highlight, and honestly, deeply validating. And then there were moments when I found myself halfway into a tech stack schematic wondering if anyone else in the room was also quietly Googling acronyms. That’s not a dig – it’s just the learning curve we’re all on.

Creative AI in action: what we heard on the ground

Some panels leaned technical NVIDIA’s Rick Champagne talked us through a comprehensive blueprint of how to use Black Forest Labs’ FLUX.1-dev as an NVIDIA NIM microservice, ComfyUI and Blender in a prebuilt workflow optimized for RTX AI PCs media workflows, for example, while others felt more like product showcases. A few gestured toward philosophical alignment on “AI storytelling” without ever defining it – which, to be fair, kind of felt like the point. Everyone’s in a different phase of experimentation, and we’re all trying to translate it into something stable.

One takeaway across multiple sessions: hybrid workflows are likely here to stay. Just as cloud computing disrupted legacy IT systems, generative tools are pushing traditional film production toward blended models. Most teams won’t go “full AI native,” but they also won’t stay fully analog either.

The Amazon MGM Studios panel, for example, emphasized moving beyond “prompt and pray.” Instead, they’re building iterative creative workflows that integrate GenAI into pitch decks, visual development, and even editorial review. The emergence of roles like “prompt artist” (people who can creatively interface with the model to produce meaningful results) underlines just how much of this shift is about communication, not just computation.

The people in the room were thoughtful, curious, and optimistic. But it’s worth noting they’re not the only ones this technology will affect. A lot of conversations still happen in the margins – in Slack threads, in hallway chats, or in blog posts like this one. So without further ado, these are my:

Five takeaways from AI on the Lot

1. There’s no one-size-fits-all approach to AI in storytelling.

The real question isn’t whether AI is good or bad. It’s whether it helps your team do better work under pressure, or adds complexity with no clear return.

2. Having the most tools won’t matter if you don’t know what you’re trying to do with them.

As access to GenAI levels out, differentiation will come from clarity of vision and execution – not tool stacks. The noise is growing, but so is the premium on specificity.

3. The best creative AI outcomes come from collaboration, not isolation.

Some of the most valuable insights came from hallway conversations. The shared friction, the honest questions, the “are you seeing this too?” moments. It’s not just about tools, it’s about the conversations around them.

4. There’s still no clear copyright framework. And there’s no proven model for profiting directly from GenAI outputs.

But in the middle of that legal gray zone, one thing is becoming obvious: infrastructure is leverage. The teams with stable pipelines, version control, and clear licensing will be the ones best equipped to scale (and defend) their use of AI. Ownership is starting to depend less on what you make, and more on how you manage what made it.

5. AI isn’t replacing creativity; it’s changing where the value sits.

At least for now, the differentiator isn’t who can generate an image. It’s who can prompt, iterate, direct, and refine in a way that feels intentional. The future of creative work will be less about “what can AI do” and more about “what do we choose to do with it?”

Where we’re focused now

We didn’t leave the conference with a single breakthrough or a perfect framework. But we left with a clearer sense of what’s required to move forward.

At KoobrikLabs, we’re focused on the unglamorous but essential foundation: infrastructure, literacy, legal clarity. The stuff that doesn’t get applause in a keynote, but that ultimately determines whether a workflow breaks or scales.

AI on the Lot was a good reminder that the tech is moving fast. But thoughtful adoption takes time, trust, and real work. We’re glad to be in that work, and we’re looking forward to what comes next.

Find out how KoobrikLabs is helping creative industries build the infrastructure and workflows behind AI adoption.