Implementing Inclusive and Reparative Metadata Practices
This workshop was part of the 2025-26 Digital Practice Workshop Series, an ongoing collaboration between the Centers for Digital Scholarship (CDS) and Humanities Center that centers topics related to digital archives and digital humanities. The series is intended to strengthen collaborations between Snell Library and CSSH.
The “Implementing Inclusive and Reparative Metadata Practices” workshop, led by Metadata and Digital Projects Supervisor Drew Facklam, gathered students, faculty, and staff for a captivating presentation and discussion on February 18, 2026. The workshop detailed strategies for detecting and correcting harmful language in archival metadata and included a demonstration of MaRMAT (Metadata Remediation and Assessment Tool).
For the past year and a half, the Digital Production Services’ (DPS) Metadata Unit has been conducting a project that focuses on developing institutional documentation on inclusive and reparative metadata. Metadata (data about data) provides descriptive language that aids the identification, discovery, and understanding of materials. While metadata may appear to be objective at times, it is deeply subjective and certainly not neutral. The understanding of metadata as subjective, and therefore incredibly capable of enacting and perpetuating harm, served as a foundational concept for both the workshop and the DPS’ inclusive and reparative metadata documentation project.
Facklam began the workshop by introducing inclusive and reparative metadata as concepts. She described inclusive metadata as “a proactive approach to representing diverse communities accurately and respectfully in new cataloging work.” While inclusive metadata is a “proactive approach,” reparative metadata grapples with existing metadata, addressing “harmful or biased description.” Together, these foundational concepts work to not only ensure that metadata refrains from enforcing and perpetuating harmful practices and stereotypes, but also centers “empathy, care, [and] respect” in the representation of individuals and communities through archival materials.
During the presentation, Facklam emphasized the importance of institutional documentation. In particular, DPS utilized the Digital Library Federation’s (DLF) Inclusive Metadata Toolkit as a grounding framework for developing accessible and informative documentation on inclusive and reparative metadata. She also described the initiation of monthly “content clubs,” where team members select and discuss readings from the DLF’s Resource Toolkit. Through initiatives such as the content club, Facklam and her team prioritize intentional collaboration in their work, as well as self- and community care.
After detailing the DPS’ specific approach to metadata documentation, the workshop highlighted best practices for projects establishing their own documentation, which included a careful assessment of current documentation and the identification of issues and gaps in existing metadata. Facklam also urged the consideration of concerns voiced by individuals and communities about their representation in collections.
A key focus of the workshop was a demonstration of the reparative metadata tool MaRMAT. Developed at the University of Utah’s J. Willard Marriott Library, MaRMAT is a free, open-source tool that employs specialized lexicons to identify and flag potentially harmful language in existing metadata. Because the tool uses a customizable list of potentially problematic terms and does not create or change content without human intervention, Facklam stressed that human decision-making is essential to MaRMAT, setting it apart from other AI tools.
The need for human-centered contributions to reparative metadata also reinforces the need for collaborative and careful documentation on workflows, resources, and processes for making metadata more inclusive and unbiased. During the demonstration, Facklam shared results of a MaRMAT analysis of AI-generated metadata on photographs from the Boston Globe Archives. The reparative metadata tool was able to identify racialized and problematic language, preparing the materials for human-led rectifications to the existing language.
Facklam and the DPS team’s thoughtful and comprehensive approach to metadata provided a useful model for project teams and individuals working with archival materials. For those who were unable to attend, a recording of the workshop in its entirety is available. If you are interested in learning more about the digital scholarship happenings at Northeastern, the CDS hosts regular events, which can be found on the CDS Events page.