The ‘Day Dreamer’ is an interactive short film that invites viewers to explore the dreams of a young Korean woman who refuses to wake up. As we delve into her dream world, her memories, desires, and fears take shape, forming a mental landscape where time dissolves and reality blurs. The viewer, by choosing which path to follow, becomes a co-conspirator in an inward escape.
The short film is composed almost entirely of images generated with Sora, OpenAI’s artificial intelligence. Only two real shots—indistinguishable to the untrained eye—anchor the film to the tangible world; everything else belongs to the language of the synthetic, the dreamed, and the unreal.
Produced by the creative agency SpecialGuestX (SGX) , a collective operating at the intersection of art, technology, and experimental storytelling. Their projects explore the limits of image and contemporary communication, blurring the lines between the real and the artificial. Among their most recent works is La Fenêtre, a documentary short about artificial intelligence currently screening at festivals around the world.
“There's something almost magical about how artificial intelligence can make transitions between any image, generating a unique aesthetic. I feel sad thinking that as Gen-AI improves, that uniqueness will be lost. I'm sure that in the future we'll miss the aesthetics of early AI, the distortions, the hallucinations.” said Miguel Espada, CEO of SpecialguestX.
“The more we’re drawn to AI-generated images, whether they feel hyper-real or utterly surreal, the more urgent it becomes to reconsider what we mean by “real.” The meaning of that word is shifting, and we’ll need to adapt to its new form. I don’t speak Korean, and I imagine most of the audience watching this film doesn’t either. But still, we accept it as real without hesitation. That tension between a real yet foreign language and familiar yet fabricated visuals produces a kind of controlled dissonance that I find fascinating.” said Lucas O. Estefanell.
With ‘Day Dreamer’, directors Miguel Espada and Lucas Ortiz Estefanell explore the narrative possibilities offered by new synthetic creation tools and open up a conversation about the futures that could emerge from technologies still in their early stages.
The images are 99% AI generated (only two shots are real) and used OpenAI Sora. The voice over has been human narrated and re-enacted with ElevenLabs.