The European Film Market (EFM) at the Berlinale became this week a barometer of a debate already running through the entire audiovisual industry: how to coexist with artificial intelligence without diluting human authorship. In hallways, panels, and marketing strategies, the discussion revealed a growing tension between technological innovation, artistic identity, and rapidly evolving business models.

One of the most talked-about gestures came from the British sales firm The Mise En Scene Company (MSC), which decided to label its market materials with an explicit seal: “No AI Used.” The company is also promoting the creation of a global standard to identify whether a film used generative AI, inspired by the precedent set by A24, which included a similar note in the end credits of the horror film Heretic. The move is not meant to reject the technology, but to establish a framework of transparency at a time when, according to CEO Paul Yates, synthetic content is beginning to flood creative industries.

“We are entering a tectonic shift,” Yates warned during the market. For the executive, human authorship risks becoming diluted if it is not defined and protected as a cultural and economic category. MSC proposes an international certification system comparable to organic or fair-trade labels, allowing audiences to understand what kind of work they are consuming. The initiative aims to bring together studios, festivals, and governments, and could extend beyond film into music, literature, and the visual arts.

The proposal reflects a broader climate of concern that also surfaced in institutional debates. This same week, the Motion Picture Association in the United States asked ByteDance, TikTok’s parent company, to stop using copyrighted works to train its new AI model Seedance 2.0, which went viral for generating hyper-realistic deepfakes of celebrities like Brad Pitt and Tom Cruise. The underlying issue is the same one hovering over the EFM: copyright, creative control, and the risk of a synthetic culture indistinguishable from the human one.

The discussions did not remain theoretical. In a session moderated by Erwin M. Schmidt, producers actively experimenting with AI brought the debate down to everyday practice. Katharina Gellein Viken, showrunner and CEO of Metrotone, and Gregor Sauter, head of emerging content at RED PONY, both pushed back against the myth of simple, cheap adoption. Integrating AI into production workflows requires sustained investment in infrastructure, expertise, and process redesign.

Metrotone, Viken explained, adopted an “AI-native” approach three years ago, while maintaining a key principle: scripts written exclusively by humans. AI is used downstream, mainly for visual generation and accelerating iteration cycles. Language models assist with research and processes, but not with idea creation. For the company, the maturity of visual tools now allows real creative control—provided workflows are designed from the outset to accommodate multiple formats.

At RED PONY, the focus is an internal “AI studio” functioning as an infrastructure layer for emerging content, especially microdramas. The system, still in beta, does not replace authorship but accelerates development: it analyzes narrative pacing, detects cliffhangers, and generates rapid prototypes for pitching. “The core value is acceleration,” Sauter summarized, while insisting that creative control must remain in human hands.

Interestingly, both producers avoided the most common narrative around cost savings. Rather than selling AI as a budget-cutting tool, they framed it as a mechanism to experiment faster, fail earlier, and reduce risk. That logic connects with another trend discussed at the EFM: the rise of vertical microdramas and the need to adapt storytelling to new consumption surfaces.

For Metrotone, projects like Raynmaker aim to rethink a format currently dominated by Asian melodramas and one-to-three-minute episodes funded through microtransactions. The strategy is to develop content conceived from the start as adaptable: a single intellectual property capable of existing as a vertical series, a horizontal feature film, and multiple spin-offs. RED PONY, by contrast, is committing fully to verticality in its new projects, aligned with specific platforms and distribution models.

The differences deepen around financing and distribution. While microdramas rely on revenue share, freemium models, and access to platform data, traditional models are still exploring hybrid schemes involving brands and creator programs. In both cases, AI enables incremental content testing, reducing upfront risk and bringing creative development closer to audience response.

The conversation closed with the concept of “liquid IP”: intellectual properties designed to flow across formats, platforms, and media. For some, this logic is inevitable in a post–peak TV ecosystem where production no longer stops at development and financing, but requires mastering technology, data, and transmedia strategy. For others, it represents a creative and financial challenge that redefines the producer’s role.

Between “No AI Used” labels, in-house AI studios, and calls for global regulation, the EFM made one thing clear: the debate is no longer whether AI will transform cinema, but under what rules. In Berlin, the market seems to have accepted that the future will be hybrid—but the battle to define its standards, and preserve the value of human creation, has only just begun.