FIAT/IFTA World Conference 2025

I present an AI-driven system for the automatic retrieval and segmentation of video content in which specific artworks are discussed. Given only the title of a work of art, the system identifies and extracts short, relevant video portions where that artwork is explicitly explained—even when it appears within broader, more general content.

The pipeline follows a multi-step process. First, I perform a keyword-based search across large-scale media archives to retrieve a ranked list of candidate videos—the top-K most likely to contain references to the target artwork. Each selected video is then transcribed using Whisper, with speaker diarization to distinguish different voices.

Next, I segment the transcription into longer monologue-style blocks, where a single speaker talks continuously for at least 30 seconds. These segments, along with the artwork title, are processed by a large language model (LLM), which identifies the portions of speech specifically related to the artwork. All original timecodes are preserved, enabling precise extraction of temporally-aligned subclips.

The output is a curated set of “shorts”—concise video segments that explain the chosen artwork—ready for use in educational, curatorial, or commercial settings. Museums can assemble engaging displays, educators can embed authentic expert commentary into lessons, and media organizations can trace and manage rights related to artwork representations across archives.

Additionally, the LLM can automatically generate relevant questions based on the content of each segment. This makes it possible to associate specific shorts with the questions they answer, enhancing both discoverability and pedagogical value within the archive.

Integrating audiovisual media with related materials in other formats in a user-accessible manner presents a significant challenge for many institutions. However, making these resources available on a unified platform can allow materials to be even more technically expressive and valuable to research. At Lawrence Livermore and Los Alamos National Laboratories this integration is particularly critical. Researchers frequently rely on both original motion picture footage of nuclear tests and the corresponding technical documentation to conduct thorough analysis and data interpretation. The labs have recently initiated efforts to address this need by leveraging a video streaming platform that allows us to collocate audiovisual materials with related textual documents, greatly improving accessibility and contextual understanding.

This poster will explore the benefits of combining diverse media formats, the challenges associated with making them accessible to researchers, and the technical obstacles involved in digitizing and preparing historic film materials—particularly those dating back to the early 1960s. As a case study, we will examine the film footage from Starfish Prime, the largest nuclear test conducted in outer space, with associated reports and documentation. This consolidation of information will provide valuable insights into the environmental impacts of the nuclear test and the unexpected effects of an electromagnetic pulse on Hawaii’s electrical grid.

This presentation focuses on a significant initiative undertaken by the Royal Danish Library in response to the Danish Media Agreement (2023-2026), aimed at enhancing public access to the archives of the Danish Broadcasting Corporation (DR). The purpose of this project is to empower Danish citizens to explore their shared history and gain insight into the events that have shaped their society.

Pivotal to the project is the establishment of a licensing agreement with the rights holders to facilitate public access to the content. Another key component of the initiative is the development of a user-friendly access platform for the DR archive, designed to allow users to search and access content without requiring login credentials. The effort includes enriching metadata to address the needs of users while ensuring compliance with legal standards. The project also entails processing DR's extensive audio and video collection spanning from 1931 to 2025, making broadcasts identifiable and accessible through segmentation and metadata enrichment processes.

To achieve these objectives, the project employs agile methodologies and integrates user experience (UX) design techniques, including focus groups and user journey mapping, to create an intuitive interface that enhances user engagement.

Ultimately, this initiative aims to preserve and provide access to DR's invaluable content, reinforcing the democratic foundation of society by ensuring equitable access to cultural heritage. This aligns with the Royal Danish Library's commitment to fostering a democratic and informed knowledge society, making this presentation a vital contribution to discussions on digital accessibility and cultural preservation.

This essay explores the evolving challenges to freedom of expression and freedom of the press in both historical and contemporary contexts. Starting with the clandestine origins of the Brazilian press in 1808, it examines the persistent tension between media censorship and the right to free speech, particularly during authoritarian regimes like Brazil's 1964–1984 military dictatorship. The discussion transitions to the digital age, highlighting the 2016 controversy where Facebook censored a Pulitzer-winning Vietnam War photo, sparking international outrage and raising questions about corporate control over public discourse. The author argues that while private platforms like Facebook wield immense power, society must actively defend freedom of expression through open debate and legal recourse. The piece concludes with a strong endorsement of democratic values, affirming the necessity of protecting the right to speak freely, even in the face of powerful interests.

As a result of the SAFE – Sustainability of Italian Film Heritage project (2022-2025), this poster will provide insights into the outputs collected from a national survey of practices and strategies for non-theatrical film heritage in Italy. The SAFE project analyzed the current state of film archiving through three research axes: storage infrastructures, digital preservation practices, and stewardship strategies.

Submitted in early spring 2025, this survey investigates several crucial topics related to sustainable film archiving practices, such as long-term analog and digital storage, labor organization, funding, and environmental accountability. The expected results of this research are twofold: first, it allows a clear and in-depth overview of the current state of preservation of non-theatrical film heritage, and second, it provides a timeline for archival organizations to anticipate the foreseeable future of preservation practices. In the last decades, alongside the more established cineteche and film archives, the field of Italian film heritage has acknowledged the rapid development and emergence of organizations dedicated to non-theatrical heritage (including home movies, political films, and industrial or scientific audiovisual production).

In this presentation, we will make available the main outputs of the quantitative and qualitative research that involved several key film heritage organizations, completed by field interviews, through a still poster including aggregated data visualization of the study. We also plan to provide an interactive data restitution of the survey results using tablets allowing cross-checked research to have a better understanding of the current state of sustainability of Italian film heritage.

This proposal explores representations of tattoo culture in the Music Television Brasil archive, focusing on the 1990s and 2000s — a period of intense cultural and aesthetic transformation in Brazil and the years during which the channel was active. It investigates how MTV Brasil’s programming and music videos catered to youth audiences engaged with alternative aesthetics and bodily expressions, especially tattoos, positioning the network as a key player in legitimizing and popularizing tattoo culture in the country.

The study is grounded in the hypothesis that, by featuring tattooed bodies—both in music videos and among its VJs—MTV contributed to the visibility and cultural recognition of a practice still marked by stigma and marginality. It introduces the concept of “tattooed gestures” to analyze music videos as an emerging audiovisual language and tattoos as bodily inscriptions that function as visual pedagogy, shaping youth identities, styles, and senses of belonging.

The core corpus will comprise over 35,000 Betacam tapes from the MTV Brasil collection, currently held by Editora Abril. Given existing restrictions and preservation challenges, the project proposes an archival intervention that combines institutional dialogue with public campaigns for digitization and access. Complementary material will be drawn from recordings available on the Internet Archive.

This study aligns with the debate on archives as interventions in problematic pasts, examining the tattoo as a historically racialized and subcultural mark. By exploring how tattooed bodies circulated on MTV Brasil and how these representations helped reshape norms around youth and embodiment, the study underscores the value of audiovisual archives as critical tools for understanding cultural change. It also highlights the underexplored significance of the MTV Brasil archive for the broader history of Latin American media and youth cultures.

Today, we have access to more data than ever before, and in the last few years, it has become widely possible to create and share artificial and potentially misleading information. Traditionally, professionals like librarians, documentalists, archivists, and information specialists helped the public access knowledge through methods like indexing and searching. Now that everyone can access, create and share data easily on their own, some might wonder if these roles are still important.

However, we argue that the sheer amount of data available to the public makes effective processing and retrieval of information more essential than ever. Similarly, the rise of artificial information calls for dedicated efforts to ensure that reliable and accurate sources of information are available to the public. In this context, documentalists can play a key role by learning new skills and adapting their existing expertise to new challenges.

This poster examines the changing skill sets that documentalists need to address these challenges, highlighting the importance of data literacy, ensuring transparency, and tracking the sources of information. We share case studies from our organization that illustrate various projects aimed at enhancing users’ access to information in this complex landscape. We discuss how improving data literacy can lead to more efficient internal processes and how teams work to enhance metadata to make information more readily usable for the public. Finally, we highlight the continued importance of traditional skills in the work of documentalists to ensure that the public have access to trustworthy sources of information in the future.