Decolonizing Libraries and Archives: Fostering Indigenous Knowledge and the Role of AI

Decolonizing Libraries and Archives: Fostering Indigenous Knowledge and the Role of AI

The discussion between Alexia Hudson-Ward and Jordan T. Clark is a thought-provoking continuation of their earlier dialogue on decolonization efforts in libraries. It focuses on collection practices, archival work, and the increasing role of artificial intelligence (AI) in academic settings. This conversation addresses how libraries can actively decolonize their spaces, collections, and practices while navigating the challenges of emerging AI technologies.

Decolonizing Libraries: Where to Begin

Clark opens the discussion by tackling one of the most frequent questions he receives: "What books should I buy to decolonize my library?" He emphasizes that decolonizing libraries is not merely about acquiring a checklist of books written by Indigenous authors but about a holistic, transformative mindset shift that inspires and motivates every aspect of library practice.

Key Takeaways:

1. Beyond Book Collections: While having literature from Native voices is essential, libraries must also physically and conceptually transform their spaces. Clark recounts an example where a librarian regularly rearranged the library to challenge students to engage with non-Western perspectives through rotating art pieces, quotes, and visual stimuli. This active engagement creates a dynamic learning environment.
2. Engagement Over Passivity: Libraries should not be passive repositories but active spaces inviting users to engage critically with their collections. For example, the librarian can reach the entire student body in high school libraries, making the library an interactive classroom.
3. Inclusive Spaces: Clark highlights Julie Fiveash's work at Harvard's Tozzer Library, which goes beyond books to include zines and other non-traditional forms of knowledge. This promotes the idea that libraries can be reimagined to offer more inclusive, non-Western perspectives.

Archives and Special Collections: Addressing Colonial Legacies

The conversation then shifts to archives and special collections, traditionally dominated by Western narratives. Clark discusses the importance of rethinking how libraries approach these collections, especially when dealing with materials historically used to oppress Indigenous communities.

Challenges and Opportunities in Archival Work:

1. The John Eliot Bible: Clark brings up the example of Harvard's Houghton Library, which houses the John Eliot Bible—the first Bible printed in the Western Hemisphere, but in the Wampanoag language, used to convert Indigenous people to Christianity forcibly. Rather than solely allowing this object to reflect a colonial narrative, it should be used to tell a fuller, more nuanced story, including its role in the modern-day Wampanoag language revitalization project.
2. Holistic Storytelling: Archivists should focus on telling the entire story of these objects, not just their colonial history. Centering Indigenous voices, such as the work of Jesse Little Doe Baird in language revitalization, helps shift the narrative away from colonial oppression toward empowerment and cultural survival.
3. Partnerships and Networks: Libraries must build coalitions with Native communities and other institutions to foster an inclusive network that elevates marginalized voices. Sharing knowledge across institutions allows for a broader understanding of how to approach decolonizing archives and special collections.

Decolonizing AI: Challenges and Cautions

Hudson-Ward and Clark then explore how artificial intelligence is both a tool for advancing Indigenous knowledge and a source of concern, as it can perpetuate existing biases rooted in colonialism.

AI in Academia and Indigenous Knowledge:

1. AI's Dependence on Existing Data: Large language models (LLMs) like ChatGPT do not create knowledge; they scrape information from existing sources. Clark notes that AI will only perpetuate those biases if the scraped data is colonized or biased. This is particularly concerning for Native communities whose knowledge has historically been stolen or misrepresented.
2. Amplifying Erasure: AI systems trained on colonized data sources could unintentionally amplify the erasure of Native voices rather than uplift them. The danger lies in reinforcing problematic narratives if Indigenous perspectives are separate from the foundational data these systems learn from.
3. Representation in AI Development: Indigenous voices must be included in developing and deploying AI technologies to combat this. Clark stresses that without Native representation in AI development, these technologies are unlikely to advance Indigenous knowledge effectively.

Building Trust and Relationships with Indigenous Communities

One of the central themes in the discussion is the necessity of building trust between institutions and Native communities. This trust is vital for fostering knowledge sharing and ensuring that Indigenous knowledge is not appropriated or misused.

Recommendations for Libraries and Academic Institutions:

1. Fostering Long-Term Relationships: Clark underscores that libraries must build consistent, trustworthy relationships with Native communities. This will allow Indigenous voices to be centered on collecting, preserving, and disseminating knowledge.
2. Representation Matters: Institutions must ensure that Indigenous people are actively involved in decision-making processes, whether in hiring Native staff, curating collections, or creating AI tools. Representation within institutions is essential for meaningful progress.
3. Mutual Benefit: For Native communities to share their knowledge, institutions must provide reciprocal benefits, such as increased educational opportunities for Native students or collaborative partnerships that benefit the community.

AI, Cultural Heritage, and the Future

The conversation concludes with reflections on how AI technologies can be both a challenge and an opportunity for Indigenous communities. Hudson-Ward mentions how African nations call for AI tools that better represent their cultural artifacts and languages, which mirrors the need for Native American representation in AI systems.

Balancing AI's Benefits and Risks:

1. Developing Culturally Relevant AI Tools: Clark emphasizes the need for AI tools to process Indigenous languages and cultural artifacts accurately. Without such developments, AI risks perpetuating stereotypes or misrepresenting Indigenous knowledge.
2. Collaborative AI Development: Universities and tech companies must collaborate with Indigenous communities to build AI systems that reflect their cultural and historical realities. This will help avoid the pitfalls of technology that misinterprets or distorts Indigenous knowledge.

Moving Forward with Intentionality

In this engaging conversation, Clark and Hudson-Ward offer valuable insights into how libraries and archives can better support Indigenous communities by decolonizing their practices and spaces. Whether through rethinking how collections are curated, building trust with Native communities, or ensuring Indigenous representation in AI development, libraries can take multiple paths to foster a more inclusive environment.

The conversation highlights that decolonization is not a checklist but a mindset and practice that must permeate every layer of an institution—from book collections to AI systems. By focusing on building relationships, fostering inclusivity, and using technology responsibly, libraries and academic institutions can begin to reverse colonialism's legacy and create spaces where Indigenous knowledge is truly valued.