Whose Bias Collective

Description

The Whose Bias Collective emerged from the "Whose Bias" event a critical intervention designed to demystify how AI and Machine Learning work and expose the inherent biases within AI systems. Born from a collaboration between digital humanists and the Youth Centre, SoapBox, we have evolved from a single event into an infrastructure for equitable research. We operate at the intersection of Digital Humanities, the heritage sector, and the third sector, challenging the traditional boundaries between "researcher" and "subject."

Our values

• Redistributing authority: We address the semantic gap between institutional expertise and lived experience. We argue that community partners hold practice-based knowledge that is equal to academic expertise and that it should be treated as such.
• Structural sustainability: We move beyond short-term projects to advocate for long-term resource sharing that preserves community organisations as the backbone of digital research and learning.
• Polyvocality: We believe technical infrastructure (from dataset schemas to interface designs) must be capable of representing diverse voices. We reject the idea that simply removing biased language solves the problem of potential historical harm; instead, we focus on pluralising records through community-led interpretation.

What we do

We co-create methods to interrogate the impact of digital technology, primarily AI, by bringing together humanities-led enquiry and community-based practice.

• Critical AI literacy and making: We co-develop hands-on workshops with young people and those typically excluded from digital policy. Young people then go on to co-facilitate events to demystify Machine Learning for the public. By using their own lived experiences, they challenge how these systems are built, drawing attention to which voices are often erased and who is marginalised by algorithmic mediation.
• Recontextualising heritage: We experiment with new ways to navigate biased or outdated language in historical metadata. Using AI tools as a starting point to flag problematic terms, we invite community members to intervene in archival descriptions. By adding community-led interpretive glosses and annotations, we enable the catalogue to reflect the lived experience of audiences and trace how meanings change over time. This intervention directly improves discovery practices by integrating the community's own vocabulary, making the archive more accessible and relevant.

Who we work with

Our collective is a cross-sector partnership:

• Academic research institutions: Providing humanities-led perspectives and technical analysis of digital systems.
• The third sector: We work with community and youth centres for on-the-ground integration and co-building research methodologies. Such partnerships allow us to amplify the voices of those typically excluded from digital policy and research.
• Cultural heritage institutions: Including museums and archives where we test our methodologies and enrich institutional records with community-grounded insights, re-evaluating historical data through lived experience, making heritage more accessible and representative. 

Currently active

Yes

Offers funding

No, this infrastructure does not provide funding.

Founding year

2025

Contact details

On the map

Go to larger version of this map

Categorisation

Type

Tags

University affiliation(s)

University of London
London

Partner Infrastructures

Warburg Institute

The Warburg Institute is globally renowned as a leading centre for the study of the interplay between ideas, images, and society. Founded in Hamburg by the eminent historian Aby Warburg (1866-1929), w… read more about Warburg Institute

London