Kling AI brought its ongoing conversation about AI filmmaking to the Toronto International Film Festival's industry market on Thursday, with a panel focused on the practical realities of blending generative video tools with human creative direction. The session, "A New Creative Paradigm: Consistent, Cinematic AI Video Powered by Kling AI," drew an audience of Canadian animation and live-action creators looking for clarity on their roles in shifting production workflows.
Moderated by Tony Pu, Kling AI's head of creator community and industry partnerships for Europe and North America, the panel featured two producers working directly with the company's platform. Diane Shorthouse, producer and co-founder of Evolutionary Films, and Issa Sissoko, AI filmmaker and founder of Amissa Studios, discussed their hands-on experience building hybrid pipelines. Kling AI had previously announced its role as exclusive global technological brand partner on Evolutionary Films' animated feature "Minibots" during the Cannes Film Market.
The script remains the foundation
Shorthouse pushed back against the notion that AI filmmaking revolves around prompt engineering. "The script is still the blueprint, so I'm not thinking about prompts, I'm thinking about shots," she said. Her process starts with the same questions any producer asks: where the camera goes, what the performance requires, and which assets are needed to execute the director's vision. The difference lies in how those assets get generated and evaluated.
"We'll generate 20 shots of something, and 19 are going to be wrong," Shorthouse said. "But those 19 are important. It's human judgment that you need to know what's wrong." That judgment, she argued, comes from understanding cinematography, lighting, and performance - not from mastering a particular software interface.
Sissoko, who has consulted for directors including Antoine Fuqua, Darren Aronofsky, and Doug Liman, described a similar dynamic when pre-visualizing scripts. His work helps producers and directors spot problems early and decide where to allocate real production budgets. "We collaborate with VFX teams as well," he said. "They might have something ready in 3D, and they want us to kind of slap the skin on it, give them an overall idea of what the cut might look like."
The control problem in hybrid production
For live-action projects that blend AI-generated elements with filmed footage, Sissoko identified a persistent technical challenge. "Getting the AI to not only keep the actor's performance intact but add to it is very tricky," he said. "That's one of the things that I work on." The goal is reducing dependence on what the model outputs by default and increasing the filmmaker's ability to direct specific, isolated changes.
Shorthouse put the same frustration more bluntly. "I need them to obey me more," she said, drawing laughter from the audience. "If I say, I want this character here in this costume in this environment but just change the eye line, I just want it to just change that not everything else." Both panelists said they expect greater control capabilities to arrive in the coming months, driven by the rapid iteration cycle they've already observed.
Sissoko pointed to the pace of improvement as reason for optimism. "Every 90 days there's a huge improvement - I mean, if you're going from three-second-generated videos to 15 seconds with more consistency across the board," he said. He noted that even current tools can serve a concrete purpose: turning a pitch deck into a concept trailer that gets a story picked up, creating work for entire crews downstream.
What emerging AI filmmakers should learn first
Shorthouse's advice to the filmmakers in the room was to build traditional craft knowledge before layering on AI tools. "Know what shots you want, how things work from cameras to lighting to performances, and then experiment with AI," she said. "And put all your knowledge of filmmaking into the prompts."
Sissoko went a step further, identifying writing as the single most critical skill for anyone entering AI-assisted production. "Before anything, learn how to write stories," he said. "It will help you with the prompts, as well." For creatives looking to build that foundation, structured training resources like an AI Learning Path for Scriptwriters can help bridge traditional storytelling craft with emerging production tools. As tools like Kling AI move from experimentation into real production environments, the panelists made clear that the underlying disciplines - script structure, shot composition, performance direction - remain the work.
Why this matters for creatives
The panel's central message was practical: AI video tools are advancing quickly, but they are not replacing the core skills of filmmaking. Producers who understand shot language, story structure, and performance will be the ones who can direct these tools effectively. The bottleneck is not the technology's capability but the creator's ability to articulate what they want and judge what the tool produces. For creatives currently in production roles, the immediate opportunity lies in learning to integrate AI-generated elements into existing workflows - pre-visualization, concept trailers, and hybrid VFX - rather than waiting for fully autonomous solutions to arrive.
Your membership also unlocks:
