Redefining the Portuguese Overseas Archives (AHU): Digital Humanities and AI for Global Knowledge Accessibility

From left: Agata Błoch, Guillem Martos Oms and Irene María Vicente Martín, Barcelona 2025.

We participated in the International Archives Congress, ICA Barcelona 27 – 30 octobre 2025, “Knowing Pasts, Creating Futures.”

This global event brought together archivists, historians, digital humanists, and professionals to reflect on how archival practices shape our understanding of the past and help build more inclusive futures.

I was especially pleased to organize our roundtable, "Redefining the Portuguese Overseas Archives (AHU): Digital Humanities and AI for Global Knowledge Accessibility". This session grew out of shared discussions about how digital tools are transforming access to historical materials, particularly those related to colonial Brazilian, and African histories. Working with Guillem Martos Oms, Irene Vicente-Martín, and Clodomir Santana, we brought together diverse perspectives to rethink how colonial collections can be made more open, interconnected, and globally accessible for researchers and communities.

Roundtable: Redefining the Portuguese Overseas Archives (AHU): Digital Humanities and AI for Global Knowledge Accessibility

Chair: Agata Błoch

Archives are using digital humanities and AI to transform preservation and access. This panel explores their use in Portuguese historical records, focusing on underrepresented regions such as South America and Africa. Using the MAPE dataset of 180,000 records, we show how NLP and generative AI reveal hidden patterns. We highlight open data, equitable digitization and the role of AI in breaking down biases and reshaping knowledge paradigms in archives.

1. Archivo en los trópicos. El fondo Dembo del Arquivo Histórico Ultramarino y el acceso a la documentación luso-africana.

   Guillem Martos Oms (University of Barcelona and National Archive of Catalonia, Spain)

En su intervención en la Conferencia Bienal de la Sección Africana Oriental y Central del Consejo Internacional de Archivos de 1982, Ali Mazrui destacó la fragilidad de la tradición archivística en África en comparación con Europa. Antes de la llegada de los europeos, solo unas pocas culturas africanas habían desarrollado sistemas de escritura y almacenamiento documental, como las civilizaciones del Nilo (Egipto, Etiopía) y las regiones del África musulmana, con ejemplos destacados como la ciudad de Tombuctú. Este contexto pone de manifiesto tanto las dificultades como los retos que plantea la preservación documental en el continente, sin dejar de señalar las oportunidades únicas que ofrece. Esta comunicación tiene como objetivo analizar las potencialidades de los archivos históricos portugueses y angoleños como fuentes clave para los estudios en Humanidades Digitales. En particular, el fondo Dembo, conservado y digitalizado por el Arquivo Histórico Ultramarino de Lisboa (AHU), se presenta como un caso paradigmático de archivos producidos por los propios africanos tras siglos de interacción comercial, diplomática y militar con los portugueses. Este fondo, que recoge documentación relacionada con las relaciones entre Angola y Portugal, ilustra la riqueza y complejidad de las fuentes documentales africanas y su capacidad para contribuir al conocimiento histórico global. A través del análisis y experiencia con el fondo Dembo, esta presentación aborda cómo las iniciativas de digitalización pueden facilitar el acceso a estas fuentes y fomentar su utilización en la investigación académica. Asimismo, se examinan los desafíos persistentes en el acceso y la gestión de estos archivos, desde la conservación hasta la democratización del conocimiento. De este modo, se pone de relieve el potencial de los archivos como puente entre el norte y el sur, mientras se exploran las posibilidades que ofrecen las tecnologías digitales para enriquecer las narrativas históricas.

2. Archives in the Tropics: Exploring Brazilian Archives Through Digital Humanities

   Irene María Vicente Martín (University of Salamanca, Spain)

Digital Humanities offers opportunities and posits challenges in the current use and access of archive documentation. This talk explores the potential of DH tools, using the Brazilian archives (Biblioteca Nacional, Arquivo Publico do Estado da Bahia and Arquivo Municipal de Salvador) as a case study, to examine the opportunities and challenges associated with integrating these methodologies into archival research. DH has the potential to enhance access, preservation, and analysis of archival materials, particularly in a context where historical resources remain endangered, poorly described or unevenly distributed. However, key challenges, such as the need for infrastructure, data standards, capacity-building, and socio-political factors, continue to influence archival practices in Brazil. An additional ethical challenge is ensuring that DH tools enhance access to archives not only for users in Europe who possess DH knowledge but also for Brazilian researchers themselves, who may face barriers related to infrastructure, digital literacy, and equitable access. The discussion will offer insights into leveraging DH tools to overcome these challenges and advance archival research in Brazil.

3. Archives & AI: Opportunities and Challenges in Digital Historical Research

   Clodomir Santana (University of California - Davis, USA)

The intersection of archival practices and artificial intelligence (AI) opens up new avenues for historical research while posing a critical challenge when it comes to combating bias and underrepresentation in digital datasets. As archives enter the digital age, artificial intelligence is emerging as a powerful tool for digitalizing, analyzing, and reinterpreting vast corpus of historical data. However, this transformation is not without its challenges, and we aim to explore them using the Portuguese Overseas Archives in Lisbon (AHU) as a case study. From the data preparation to the section and application of the model, we demonstrate the potential of techniques such as Topic Modelling, Natural Language Processing, and Social Network Analysis. We discuss issues related to preprocessing, data scarcity, the reliability of AI models, and potential data biases that reinforce the marginalization of narratives of underrepresented groups, cultures, and languages. We also address strategies to mitigate these barriers, including using transfer learning, low data models, data balancing, and usage of culturally specific datasets. This paper positions AI as a transformative force in digital history and acknowledges that its success also depends on good archival practices and digital data accessibility.

4. Cracking the Historical Code: Big Data and Computational Models with the MAPE Dataset

   Agata Błoch (Institute of History, Polish Academy of Sciences, Poland)

The vast scope and complexity of historical archives present researchers with major challenges, particularly the overwhelming amount of data and its often unstructured nature. This presentation shows how large language models (LLMs), combined with data science techniques, can transform the way scholars explore such vast collections. Using the MAPE search engine as a case study, applied to our dataset of over 180,000 correspondence entries from the Portuguese Overseas Archives in Lisbon (AHU), dating from 1610–183, we show how LLMs can replace traditional keyword-based searches with natural language queries. This shift allows researchers to navigate intuitively through archival content and gain insights that would be difficult to access using traditional methods. Finally, we reflect on the practical and transformative potential of computational modeling in digital history. While challenges remain particularly in managing scale, complexity, and interpretation, our experience shows that digital tools make it possible to navigate, analyze, and interpret vast archival collections. By combining the richness of historical archives with the analytical capabilities of modern data science, we are opening up new ways of understanding historical phenomena in a way that is both rigorous and rich in context.

More information: Programme - ICA Barcelona 2025