AI film project highlights reliance on human creativity despite generative tools

Higgsfield, an AI filmmaking platform, released 'Cully Hill Boys,' an AI-generated film created using its Cinema Studio suite and the Seedance 2.5 video model. The project was designed to demonstrate the capabilities of Higgsfield’s tools, including its ability to generate longer video clips (approximately 30 seconds) compared to other AI video models, which typically produce only a few seconds of footage at a time.

The film’s narrative was written by human screenwriter Tim Planagan, who retained the rights to produce the story traditionally despite licensing the AI production rights to Higgsfield. Planagan does not use AI tools in his writing process, and the script’s structure—including character development and pacing—reflects human storytelling conventions. Directors Adilet Abish and Aitore Zholdaskali, alongside a team of nine additional directors, closely followed Planagan’s script to maintain narrative cohesion, a factor the reporter notes as a key reason for the film’s relative strength compared to other AI-generated projects.

The film’s human avatars were created using photos and audio clips provided by real people, including streamer Mikyle 'N3on' Rafiq, content creator Matt Kiatipis, and UFC fighter Israel Adesanya, who licensed their likenesses for use in the project. While none of the actors performed on a physical set, their licensed digital representations contributed to the film’s visual and emotional authenticity.

Higgsfield did not pursue a traditional theatrical release for 'Cully Hill Boys.' Instead, the company uploaded the film and many of the prompts used to generate it online as a promotional showcase for its AI filmmaking tools, including the newly added Seedance 2.5 functionality. The project also highlights Higgsfield’s emphasis on collaboration between studios and digital creators, positioning the film as an advertisement for its broader suite of AI tools.

The reporter notes that despite the film’s polished appearance in some scenes, inconsistencies—such as incomprehensible text in books—reveal the limitations of current AI video generation technology. These flaws underscore the ongoing need for human oversight in AI-generated media to ensure coherence and quality.