Directing the Invisible: AI, Memory, and Rhythm in HEJJAYTI Pt. VII

Introduction

As a Moroccan multidisciplinary artist and Modern Griot, my creative practice lies at the intersection of oral storytelling, sound design, and experimental cinema. I grew up listening to stories whispered in marketplaces, sung in sacred ceremonies, and carried across the air of my city. These sonic memories became the foundation of my artistic identity. In an era dominated by rapid technological innovation, I have sought to explore how ancestral voices and artificial intelligence can coexist and weave together new narrative forms.

HEJJAYTI Pt. VII emerged from this inquiry as a hybrid experiment that reimagines authorship, identity, and rhythm in the digital age. The song, composed and performed during my first year of the Master’s in Documentary Cinema, is rooted in Moroccan musical traditions and personal introspection. The video, however, ventures into a speculative visual realm made possible through advanced AI tools. This combination of tradition and innovation reflects my belief that technology, when approached critically and poetically, can act as an amplifier of cultural memory rather than a threat to it.

Methodology

The creation of HEJJAYTI Pt. VII unfolded through two intertwined processes: the composition of the song and the development of its visual narrative using AI. The musical component draws on Moroccan rhythmic patterns and contemporary spoken-word poetics. The lyrics are intimate, reflecting on the weight of memory and the desire to break free from cycles of silence.

For the visuals, I initially experimented with Midjourney to generate reference images. However, this tool struggled to produce authentic representations of my face and emotional nuances. DreamBooth, built on Stable Diffusion, became the cornerstone of the workflow. By training the model with curated images of my face and expressions, we achieved a level of realism that preserved identity and artistic intent. This process demanded hours of trial and error, fine-tuning prompts and parameters until the AI could mimic the subtlety of a gaze or the texture of skin.

Once the static images were perfected, Kaiber AI was employed to animate them into fluid sequences. These animations transformed still portraits into dynamic visuals that pulsed with the rhythm of the music. The final result is a video that resists linear storytelling, instead unfolding as a visual poem synchronized with the soundscape.

Case Study: HEJJAYTI Pt. VII

HEJJAYTI Pt. VII began as a deeply personal composition, an act of resistance against the noise of contemporary existence. Its lyrics oscillate between introspection and defiance, and the vocal delivery draws inspiration from traditional forms like Moroccan music.

Visually, I wanted to translate these emotions into a world that felt both real and imagined. The initial AI experiments were frustrating: Midjourney often failed to produce faces that resembled me, and when it did, the expressions felt hollow. DreamBooth provided a breakthrough. Training the AI to recognize and reproduce my facial features allowed me to reclaim a sense of agency in a process often critiqued for erasing identity.

The script for the video was meticulously constructed, with each scene guided by detailed prompts that blended Moroccan cultural motifs with surreal elements. For example, one sequence depicts a man with a full beard flying over the Sahara desert, letters falling from the sky like snow. Another shows vibrant Moroccan streets where paintings come alive, their brushstrokes animated as if infused with hidden energy. These scenes, generated through DreamBooth and animated via Kaiber, reflect a negotiation between my artistic vision and the machine’s interpretations.

AI and the Modern Griot

As a Modern Griot, I see myself as a keeper of stories and a weaver of sound. In my practice, AI becomes part of the griot’s ensemble, an unseen musician contributing new textures to the collective memory. This perspective challenges dominant narratives around AI as a threat to artistic integrity. Instead, it proposes a methodology where artists in North Africa and beyond can use technology to amplify marginalized voices and preserve cultural heritage in innovative ways.

Conclusion

HEJJAYTI Pt. VII demonstrates the potential of AI to expand the language of cinema and sound. This work positions artificial intelligence not as a replacement for human creativity but as a collaborator capable of generating new narrative forms. It reflects a commitment to blending tradition with innovation, offering a roadmap for artists seeking to engage with technology without losing cultural grounding.

In the context of African and North African cinema, this experiment opens doors to storytelling practices that are deeply local yet globally resonant. It invites us to imagine a future where machines do not overwrite memory but help us listen more deeply to its echoes.

Bibliography

Ettalabi, El Mehdi. 2024. Celebrating One Year of HEJJAYTI Pt. VII. Blog post. https://mehdayevbrayen.com/blog/celebrating-one-year-of-hejjayti-pt-vii/
https://drive.google.com/file/d/1YoZaJ29Le0lc2HgT3WxTa1aYZwUSrrPS/view

Kaiber AI. https://kaiber.ai

DreamBooth (Stable Diffusion). https://stability.ai

Midjourney. https://www.midjourney.com

Kaiber Official. 2023. “How to Create AI Music Videos Using Kaiber AI.” YouTube video. https://www.youtube.com/watch?v=7NK_JOkuSVY

Filmography

HEJJAYTI Pt. VII. (2024). Directed by El Mehdi Ettalabi aka Mehdayev Brayen. Morocco.