Prix
The Prix Ars Electronica Archive is a collection enabling search and viewing of all the submissions since 1987. The award-winning projects are documented with catalogue texts and audio-visual media. All other submissions can be searched by title/artist and displayed with year, category in list form. Please cite the credits (artwork name, artist and photographer) and only use the materials if your article is related to Ars Electronica.
AI’s Silent Omissions
AI's Silent Omissions is a documentary work that examines the structural exclusion produced by contemporary image-generation technologies, in which only bodies defined as "normal" are selected, while others are treated as outliers and systematically removed.
In AI-generated imagery, hands that do not conform to the five-finger standard are often labeled as errors or defects and discarded. Engineers devote vast computational resources to refining algorithms so that AI can more accurately reproduce a standardized human body. Yet in this pursuit of technological "progress," the lives and bodies of minorities—such as people with cleft hand—are rendered invisible, quietly erased from both datasets and imagination.
The film centers on the cleft hand as a symbolic motif. Through a series of interviews conducted by an AI engineer with affected individuals, their families, and medical professionals, the project unfolds as a process of dialogue and self-examination. As these conversations progress, the engineer's own implicit assumptions about normality and standardization are gradually exposed, revealing the biases embedded in technological systems.
The work extends beyond questions of bodily representation. It raises an ethical inquiry into how databases describe the world: what they choose to make visible, and what they render invisible. The voices, forms, and bodies treated as "noise" by AI systems nonetheless carry irreducible dignity and beauty.
AI's Silent Omissions is at once a poetic documentary that seeks to recover these marginalized presences and a critical examination of the mechanisms of classification and exclusion that shape contemporary society. By placing technology in direct relation to human diversity, the work invites reflection on what is lost when optimization becomes the dominant logic of the digital age.
Links: https://nozomukubota.com/en/works/ais-silent-omissions/, https://vimeo.com/1139619987/684e3d513d?share=copy&fl=sv&fe=ci
Director: Nozomu Kubota
Cinematography and editing: Yuto Kurokawa
Participants: Yuki Asahara, Yuu Otsuka, Slime, Hidehiko Kawabata
Special Thanks to:
Professor Furukawa Kiyoshi (Tokyo University of the Arts)
Professor Nakamura Masato (Tokyo University of the Arts)
Project Associate Professor Kuribara Yoshiaki (Tokyo University of the Arts)
Assistant Nagashima Yui (Tokyo University of the Arts)
With the support of Shakimen
Nozomu Kubota (JP), born in NY in 1984, lives in Tokyo. He founded his company while studying at Keio University at the age of 19. He holds an MFA from Tokyo University of the Arts, Department of Inter-Media Art. He holds 20 AI-related patents across Japan, the US, China, and Hong Kong. With an engineer's insight into AI, he deliberately subverts its errors to illuminate marginalized communities at risk of erasure.
In AI’s Silent Omissions, Nozomu Kubota exposes a quiet violence: the systematic erasure of bodies defying coded norms. By pairing forensic dataset audits—including bypassing NSFW filters—with interviews with people living with cleft hands, the work interrogates the politics of visibility and digital dignity. It shifts bias debates from abstract ethics to lived embodiment, showing how AI misreads difference as error or obscenity. The jury commends its refusal of techno-solutionism. Employing a “reflexive methodology,” Kubota, an AI engineer, exposes his own implicit biases during the process, critiquing both the system and its creator. AI’s omissions are not empty; they are social decisions sedimented in code. The project speaks with rare clarity, restoring human density to those the algorithm sought to erase.