Mexico Examines AI to Preserve Cultural Archives

Mexico is using AI to explore new ways to classify, index, transcribe, access, and preserve sound, photographic, audiovisual, and textual archives, while experts warn that AI systems must be developed and applied ethically to prevent historical biases from shaping which parts of cultural memory remain visible.

The preservation of cultural memory is moving into a new technological phase as archives across Ibero-America examine how AI can help protect and provide access to sound, photographic, audiovisual, and textual collections. The discussion is taking place at the “Ibero-American Archives Meeting: Artificial Intelligence in Document Preservation,” organized by Mexico’s Ministry of Culture in collaboration with the General Secretariat of Ibero-American States (SEGIB) and regional archival programs.

The meeting brings together specialists from 11 countries to examine how AI can be incorporated into archival work while addressing the risks associated with its use, including bias, data protection, access to information, and the representation of historically marginalized voices. Lucina Jiménez, General Director of Cultural Training and Management, Mexico’s Ministry of Culture, says the region is entering a critical conversation about both the potential and risks of AI as archives move further into digital management. 

“Mexico begins a route of Ibero-American conversation with this second meeting, from which 11 countries come together to generate a fundamental conversation,” says Jiménez.

The importance of the discussion goes beyond adopting AI. As the experts note, archives contain records that document the memory of societies, including materials related to communities and voices that have historically received limited representation.

AI Can Make Large Archives More Accessible

One of the clearest applications discussed at the meeting is the use of AI to manage the volume and complexity of archival collections. Lorena Larios, Ambassador and Secretary for Ibero-American Cooperation, says AI can facilitate classification, indexing, transcription and access to documents. These functions can help institutions organize collections and make their contents easier to locate and consult. For archives containing sound recordings, photographs, audiovisual material and documents, such capabilities can support access to information that might otherwise remain difficult to identify or retrieve.

But the same systems that help determine what information can be found can also influence what information becomes visible. Larios says that when AI models are trained with historical collections or datasets containing biases, they can reproduce or perpetuate systemic prejudices. "That bias can determine what information will be visible in the future,” she says, adding that it can reproduce inequalities and exclusions already present in the documents being preserved.

For archival institutions, this makes the deployment of AI different from a conventional digitization exercise. The technology can accelerate information management, but its results still depend on the data and historical records used to develop and operate the systems.

Preserving Memory Requires Human Oversight

Francisco Javier Rivas, Head, Mexico’s Fonoteca Nacional, says Ibero-America should not limit itself to becoming a user of technologies developed in other countries and contexts. 

For Rivas, developing regional capabilities is particularly important when the systems are applied to collections that form part of the memory of Indigenous peoples. “The Ibero-American region cannot limit itself to being only a user of technologies that have been designed in other countries and contexts,” he says.

Rivas argues that the development of its own capabilities and the ethical and responsible use of AI should accompany Mexico’s National AI Plan. The objective, he says, should be technological sovereignty and ensuring that AI serves the public interest and people rather than becoming an end in itself.

That approach places archival professionals alongside technology as part of the preservation process. AI can assist with repetitive and information-intensive tasks, but the decisions surrounding access, representation and the treatment of sensitive collections remain central to the work of institutions and information managers.

Mariella Cornejo, President, Ibero-American Network of Diplomatic Archives, says the technological transition also provides an opportunity to reconsider the relationship between technology and human work in archival management. She says professionals, archivists, social groups, and cultural managers can use the discussion to develop another perspective on technological progress and its relationship with the organization and protection of cultural memory.

Bias Becomes an Archival Preservation Issue

The meeting also places inclusion and gender at the center of the discussion. José Vilches, Deputy Director, National Archives of the Dominican Republic, says institutions share the responsibility of protecting the memory of their societies while innovating in the methods used to guarantee its access, preservation and dissemination. He says AI introduces significant ethical challenges because its development and application must account for historical and structural biases.

The concern is particularly relevant when archival systems determine how information is classified, indexed or retrieved. If existing omissions and imbalances are reproduced through automated processes, technology could influence how future generations encounter historical records.

The encounter therefore addresses access to information, protection of sensitive data, preservation of collections and the inclusion of voices that have been minimized or silenced. Jiménez says the discussions are also intended to establish new concepts around cultural heritage and strengthen practices in cultural and artistic management.