Artificial intelligence doesn’t have to be cold or dehumanized. The project “Art, Community, and AI: Images for an Affective Memory” delves into a fascinating frontier: how AI can be used to create, reconstruct, and reimagine visual archives that never existed—allowing communities to restore emotional and identity-based memories erased by violence, time, or historical exclusion.
The study emerges from an urgent need in regions such as Latin America, where communities have lost visual records due to dictatorships, forced displacement, political repression, or simply because they never had access to photographic documentation.
Generative AI opens the possibility of producing images that, while synthetic, hold deep symbolic and emotional meaning.
Reconstructing the Past with AI
Four Latin American projects demonstrate that AI, far beyond its commercial applications, can become a powerful tool to restore fractured memories—particularly in communities that have endured historical violence, forced displacement, or the loss of visual records.
The four initiatives examined in this study—An Inexistent Queer Archive; Exhuming Memor.IA; IAbuelas, and Synthetic Memories—demonstrate how technology can serve as a techno-aesthetic device, reshaping how society perceives and feels about the world.
In An Inexistent Queer Archive (2022), Chilean artist Felipe Rivas San Martín crafts a fictional photographic archive of queer life in early 20th-century Latin America using IA.
One of the consequences of heteronormative culture is the absence of any visual record of queer experiences from the past. The generated images portray gay and lesbian couples and gender-diverse people in moments of intimacy and affection—filling a painful historical void.
Synthetic Memory and Real History
The project Exhuming Memor.IA (2023), by artist Rogelio Séptimo, takes the idea of AI-driven memory reconstruction into the community sphere. On Janitzio Island in Michoacán, Mexico, he combines ancestral Purépecha practices—such as barter tables—with artificial intelligence to visualize ancestors for whom no photographs exist.
One of the most moving examples is IAbuelas (2023), created by advertising professional Santiago Barros. The project uses AI to imagine what the babies stolen during Argentina’s dictatorship (1976–1983) might look like today. Drawing on photographs of their disappeared parents, the system generates updated portraits of these now middle-aged adults—people who, if alive, would be between 45 and 50 years old.
Importantly, these images are not intended to replace the scientific identification methods employed by the human rights group Abuelas de Plaza de Mayo. Instead, they open a space for reflection and visibility, contributing to the search for roughly 300 grandchildren who remain unidentified.
Meanwhile, Synthetic Memories (2023), developed by the Barcelona-based studio Domestic Data Streamers, focuses on memory reconstruction for people with Alzheimer’s or displaced communities. Through personalized interviews, oral testimonies are transformed into images that help participants reconnect with their past.
Layers of Memory Work
The research identifies three distinct levels of intervention in fractured memory:
- Socio-political. Projects like An Inexistent Queer Archive and IAbuelas address historical invisibility and state violence, producing counter-narratives that challenge dominant discourses.
- Community-cultural. Exhuming Memor.IA ties contemporary technology to ancestral practices to recover cultural traditions.
- Therapeutic-personal. Synthetic Memories emphasizes individual restoration, particularly in contexts of cognitive decline.
.
These projects also face significant ethical challenges. How can synthetic images avoid being mistaken for real historical evidence? Each artist has adopted specific safeguards: IAbuelas explicitly states that its portraits have no forensic value, while An Inexistent Queer Archive deliberately leaves technical imperfections to signal their artificial nature.
The research also introduces two new concepts: algorithmic postmemory—a framework for understanding AI’s active role in shaping both individual and collective memory—and affective symbolic documentary, which highlights the testimonial value of AI-generated images.
The findings suggest that a humanistic perspective and ethical commitment using AI can play a meaningful role in emotional repair, memory reinvention, and the transformation of established structures of feeling.
AI does not replace human memory; it complements it, enabling new forms of symbolic preservation and restoration. This opens promising paths for addressing historical trauma and critically examining the power structures that control collective memory.
As these projects show, the future of affective memory is closely tied to AI’s potential to help build more inclusive and reparative histories.
.
References
- Bañuelos Capistrán, J., Zavala Scherer, D., & Lugo Rodríguez, N. (2025). Art, community and AI: Images for an affective memory. Frontiers in Communication, 10(Visual Communication), Article 1567694.
- Criales, J. P. (2023, July 30). La inteligencia artificial imagina cómo se verían hoy los bebés robados por la dictadura argentina [Artificial intelligence imagines how babies stolen by the Argentine dictatorship would look today]. El País.
- Domestic Data Streamers. (2024). Synthetic memories research.
- Séptimo, Rogelio. 2023. Exhumar la Memor.IA.
Autor
Jacob Bañuelos Capistrán. Research professor in the Department of Media and Digital Culture at the School of Humanities and Education, Tecnológico de Monterrey. He holds a PhD in Information Science from the Complutense University of Madrid and is a Level 2 member of Mexico’s National System of Researchers (SNI).
