FIAT/IFTA World Conference 2026

In the high-pressure environment of major broadcasting, the demand for archival media is defined by speed, precision, and instant availability. At Globo, managing decades of rich audiovisual heritage across Journalism, Sports, and Entertainment requires a constant evolution of archival strategies. The Globo Mídia project represents the latest strategic leap in this journey, designed to amplify the reach and agility of our media ecosystem through advanced data structures.


The core of Globo Mídia lies in the implementation of ontologies that structure vast amounts of media data, providing a unified semantic language across the company. This framework provides the necessary context for AI-driven initiatives, ensuring that technology acts as a powerful ally in metadata extraction and information retrieval. By creating these intelligent connections, Globo Mídia ensures high-speed discovery and accuracy, even in the most demanding and time-sensitive production scenarios.


The ultimate goal of this technological advancement is to liberate the researcher’s eye. By using the Globo Mídia ecosystem to automate mechanical indexing and repetitive tagging, we empower our specialists to transcend labor-intensive tasks. This shift allows the human researcher to focus entirely on their role as a strategic content partner and curator, applying historical sensitivity and deep creative insight to the development of new products.


At Globo, we believe that technology should handle the scale and complexity of the data so that human intelligence can drive the storytelling. We will demonstrate how Globo Mídia bridges the gap between massive archival collections and high-demand production needs, ensuring that our historical assets are instantly accessible to fuel future creativity and narrative excellence.

Early 20th-century archaeological expeditions produced a rich body of visual material: film footage, field photographs, lantern slides documenting landscapes, excavation methods and cultural contexts. Dispersed across museums, university libraries and national film archives, these film collections remain difficult to discover, cross-reference or integrate into contemporary research. This presentation reports on an ongoing project that applies AI techniques to support descriptive cataloguing and access in digitized early film archives.
The project centers on a corpus of records from institutions in Europe and North America. Using vision language models, it explores how AI-assisted workflows help integrate contextual metadata, including site, estimated date ranges, artefact types. The presentation goes through opportunities and challenges of existing models trained on modern visual conventions that differ from historical image quality, aesthetic and stylistic composition. It observes how black-and-white historical footage with inconsistent framing and labels requires domain-specific approaches, and which solutions can be envisaged to preserve the richness of early audiovisual information.
Beyond cataloguing, the presentation examines how automated archival description affects and transforms cultural memory. Text matches and visual similarity now allow researchers to reconstruct early film sequences across archives and curate thematic collections that reframe and reposition our understanding of historical events. The paper reflects critically on the digitization of early footage and its entanglements with automation: which screen memories can machines unlock?
By sharing epistemological reflections on early documentary film culture, this presentation aims to contribute to broader conversations in the FIAT/IFTA community about the responsible integration of AI into archival workflows and the renewed relevance of historical film collections as forms of cultural mediation.

During the 2025 Club World Cup, Globo’s Archive department, in partnership with the Sports Editorial team, conducted a proof of concept (PoC) using Google Vision Warehouse to explore new possibilities for content discovery and usage in live production environments.
 
The initiative focused on evaluating how artificial intelligence can transform the way content is retrieved and utilized in sports journalism and live broadcasting.
 
By leveraging automated video processing, the solution was able to interpret specific contexts and situations within the footage. This enabled a shift from traditional keyword-based searches to a contextual approach, significantly expanding the scope of content discovery.
 
In practice, the Sports team gained the ability to access content through contextual queries—such as specific game situations, crowd behavior, or visual elements—even when this information had not been previously indexed in the metadata. This new model increased the speed of content retrieval, added depth to editorial coverage, and enabled the discovery of previously underutilized assets.
 
The PoC demonstrated the strategic potential of AI within the Archive, making content more accessible, reusable, and relevant across diverse editorial contexts.

This presentation reflects on how a practical archival task inside a Uruguayan television broadcaster evolved, over sixteen years, into a knowledge-driven system for managing, contextualising and reusing audiovisual memory.
The project began in 2010, when I was assigned the recovery of a historical television archive using only handwritten paper notebooks as the primary access tool. At the same time, much of the collection still existed on ageing analogue and early digital videotape formats (1-Inch Tape, U-Matic). To make the material accessible, I had to learn through experimentation how to clean, restore and digitise playback machines that were already several decades old.
Coming from a generation that experienced both analogue video culture and the first stages of digital editing, I combined broadcast operational knowledge with self-taught experimentation. Early attempts using Word documents and spreadsheets quickly proved insufficient for the speed and precision required in daily television production. This operational pressure led me to begin learning programming and databases in order to recreate the logic of the paper workflow in digital form: locating content quickly, identifying the correct tape, and later the precise timecode inside the material, find actors etc.
As newsroom demands increased, the system evolved continuously through real-world use, driven by requests from colleagues who relied on rapid access to historical footage. Even after the archive department was closed in 2020, the project continued independently, incorporating long-term preservation logic, production workflows, contextual metadata, artificial intelligence support and large-scale efficiency testing using a personal dataset of more than one million media assets collected over years of archival practice.
This case study explores how necessity, situated learning and broadcast reality can generate technological innovation from within archival work itself.
That's my story.