A 75-minute film about Iranian civilians killed during protests is premiering at the Tribeca Film Festival. It was made for approximately $2,000. Every frame was generated using artificial intelligence. These three facts, taken together, describe something that would have been technically impossible for an independent filmmaker just a few years ago, and they raise questions that go well beyond the usual debates about AI in Hollywood.
The film is called Dreams of Violets. Its director is Ash Koosha, a filmmaker originally from Tehran who left Iran in 2009. He made it with his brother, Pooya.
The Conditions That Made This Film Exist
To understand why Dreams of Violets matters, it helps to understand what it is responding to. According to the Human Rights Activists News Agency, at least 7,000 people died during protests that broke out in January, and more than 50,000 were arrested. The people inside Iran who witnessed these events cannot speak publicly about them. Koosha, living outside the country, chose to respond with a film.
The constraint was not creative. It was material. Producing a conventional narrative film about recent events in Iran, with actors, locations, and a crew, would require resources, access, and time that simply did not exist. AI tools for video generation, language editing, research, and imagery made it possible to compress that process into three months and roughly $2,000.
The result is a dramatized account centered on five characters, including a boy named Amir who has cerebral palsy, a woman pressured by her family to stop going out, and a man falling from a building amid smoke bombs and crowds. The film ends with five people facing execution in an alley as Amir watches. The trailer, at 83 seconds, shows smudged backgrounds that signal AI generation, but the overall visual register is coherent enough to read as a realistic film.
This is what most coverage of AI-generated media misses: the technology is not just a cheaper version of existing production pipelines. It is, in certain circumstances, the only pipeline available.
The Ethical Weight That Comes With the Tool
Koosha addressed the central discomfort directly in a public statement. Making an AI-generated film about people who actually died is not a neutral creative choice. It involves representing real suffering through synthetic imagery, giving visual form to events that the filmmaker did not witness firsthand and that survivors cannot safely describe.
His stated reasoning is worth examining carefully. He frames the alternative, silence and forgetting, as worse than the ethical risks of representation. He describes the film as existing because, in his words, the dead deserve to be witnessed and the families inside Iran, who cannot speak, deserve someone outside who refuses to forget.
This is not a technical argument. It is a moral one, and it is the kind of argument that AI tools are increasingly forcing into the open. The question is no longer only whether AI can generate convincing imagery. It is whether that imagery should exist, who has the standing to create it, and what obligations come with using synthetic media to represent real human loss.
Jane Rosenthal, who cofounded Tribeca, described the film as a powerful example of how AI can function not simply as a tool of innovation but as a vehicle for deeply human storytelling. That framing is significant. It positions AI not as a replacement for human intent but as an amplifier of it, a way for a filmmaker with a story to tell and almost no resources to tell it anyway.
Why This Moment Is Larger Than One Film
Dreams of Violets arrives at a moment when the film industry is actively negotiating its relationship with AI. The Academy has introduced new restrictions on how AI can be used in films submitted for Oscar consideration. The recent digital reconstruction of Val Kilmer in As Deep as the Grave has reopened debates about consent, likeness, and the boundaries of posthumous representation.
These conversations tend to focus on commercial cinema, on studios, budgets, and intellectual property. What Dreams of Violets introduces is a different axis entirely: AI as a tool for political witness, used by someone with no studio backing, no crew, and no access to the place he is documenting.
This matters because it changes the population of people who can make films. Documentary and narrative filmmaking about conflict, repression, and human rights has always been constrained by access and money. AI does not eliminate those constraints, but it compresses them significantly. A filmmaker in exile, working with a sibling and a few thousand dollars, can now produce something that reaches a major international festival.
That is genuinely new. It does not resolve the ethical questions. It makes them more urgent, because the barrier to entry has dropped far enough that those questions will now arise far more often, in far more contexts, made by far more people.
In Short
Dreams of Violets is not primarily a story about AI capabilities. It is a story about what happens when those capabilities meet a filmmaker with something urgent to say and almost nothing to say it with. The film cost around $2,000 and took three months to produce. It documents events in which at least 7,000 people died. The technology made it possible. The ethics of using it remain genuinely open. Both of those things are true at the same time, and that tension is exactly what makes this film worth paying attention to.
Based on reporting from Rolling Stone - AI.