March 2026

AI Cinema and the Festival Ecosystem
Some Curatorial Observations

By Enrico Vannucci

Over the past few years, generative artificial intelligence has become one of the most widely discussed developments in the audiovisual field. Text-to-video models, image generators, and other AI-driven tools are frequently presented as heralding a profound transformation of cinema itself. In public discourse, it is increasingly common to encounter the idea that “AI films” are about to redefine the future of filmmaking.

Yet the reality within the festival ecosystem appears, at least for the moment, considerably more nuanced. From the vantage point of someone working within a major international short film competition receiving around four thousand submissions each year, works making visible use of generative AI remain surprisingly rare. Among the thousands of short films submitted annually, projects that clearly rely on AI-generated imagery amount to only a few dozen — a very small fraction of the overall landscape.

This marginal presence is often interpreted as evidence of institutional resistance. Film festivals, the argument goes, remain structurally conservative and reluctant to embrace technologies that challenge established notions of authorship, labour, and artistic practice. While this interpretation contains an element of truth, it tells only part of the story.

A second factor may be equally significant: the relationship between generative technology and cinematic form is still largely unresolved. Many works that employ AI today function less as fully realised films than as experiments, visual studies, or technological demonstrations. In numerous cases, AI-generated imagery appears only as a fragment or stylistic device within otherwise conventionally produced works — an effect, a transition, or a brief visual sequence — rather than as the foundation of a coherent cinematic language. In many respects, filmmakers are still trying to understand what AI-generated imagery can actually be used for within the grammar of cinema.

At the same time, a number of filmmakers have begun to explore the opposite direction: rather than concealing the limitations of generative systems, they deliberately foreground them. Glitches, visual inconsistencies, and unexpected artefacts become part of the aesthetic strategy itself. The resulting works often adopt a playful or ironic tone, inviting audiences to recognise the machine’s imperfections. Yet these films tend to belong more to the realm of experimental or art-house practice than to any emerging mainstream cinematic model.

In discussions about new technologies, such enthusiasm is hardly unprecedented. The arrival of virtual reality roughly a decade ago was accompanied by similar claims that immersive media would fundamentally transform cinema. Film festivals quickly created dedicated sections and exhibition spaces, yet the traditional cinematic form remained largely unchanged. Rather than replacing cinema, VR simply established itself as a parallel practice within the broader audiovisual ecosystem.

Generative AI may follow a comparable trajectory, although its implications remain uncertain. The central question is therefore not merely technological but creative: how can these tools be used to produce works that function convincingly as cinema? At present, many AI-generated images circulate widely across online platforms, attracting millions of views through novelty or spectacle. Yet the popularity of such content rarely translates into works that possess the artistic coherence or formal ambition typically associated with cinema.

In this sense, the challenge facing filmmakers is not simply to incorporate AI into existing workflows, but to understand how these systems might contribute to the emergence of new cinematic forms. The short film format may prove particularly well suited to this process. Throughout the history of cinema, short films have frequently served as a testing ground for emerging technologies and aesthetic approaches. Animation, digital compositing, and early computer-generated imagery all developed through periods of experimentation in short formats before becoming integrated into larger industrial practices.

If generative video is to become a meaningful component of contemporary cinema, a similar phase of exploration will likely be necessary. Rather than appearing immediately as a fully formed cinematic language, AI may first find its place through a gradual process of formal discovery and creative adaptation.

Generative cinema may ultimately emerge not through technological breakthroughs alone, but through the gradual process by which filmmakers learn to think cinematically with these new tools.