The Bibliothèque et Archives nationales du Québec (BAnQ) is embarking on an ambitious project to develop a comprehensive database that focuses on Quebec’s cultural and governmental content. This initiative aims to enhance the capabilities of artificial intelligence systems, particularly in their understanding of Quebec society, culture, and Indigenous languages. Following a successful feasibility study, BAnQ has entered the experimental phase of this project, which is poised to address the significant gaps in data available for training AI systems.
Addressing Data Deficiencies
The necessity for this project stems from the observation that many generative AI systems struggle to accurately represent Quebec’s unique societal landscape. Valérie D’Amour, who spearheaded the feasibility study, emphasised that “all scenarios are a little bit on the table right now.” This flexibility reflects BAnQ’s commitment to engaging stakeholders from cultural sectors and data ownership groups to explore viable pathways for the database.
BAnQ’s initiative is not merely an academic exercise; it is a response to a pressing need. Current AI models often rely on a limited scope of data that fails to encapsulate the rich tapestry of Quebec’s cultural identity. Marie Grégoire, the institution’s president and CEO, articulated the goal succinctly: “We want AI systems to better reflect Quebec society and culture.”
A Controlled Data Environment
Importantly, the future database is designed to maintain strict access controls, ensuring that it does not serve as a public distribution channel for creative works. This approach aims to safeguard the intellectual property rights of creators while providing a structured framework for the integration of local data into AI systems.
The project draws inspiration from similar initiatives, such as one in Sweden, which has successfully assembled extensive collections of Nordic-language texts to train generative AI models. BAnQ plans to begin with its own collections, gradually considering external data sources as the project evolves.
Cultural Representation and Ethical Considerations
The drive to create this database is underpinned by a broader concern regarding the representation of Quebec culture in AI. Destiny Tchéhouali, a prominent figure in research on French-language AI, highlighted the risks of cultural and linguistic biases within existing AI datasets. “Quebec culture remains underrepresented in the corpora currently circulating in the AI world,” he noted.
Tchéhouali advocates for the database as a “strategic infrastructure” that could provide essential guidelines for identifying, cataloguing, and tracking local content. This proactive approach is crucial, especially in light of the potential for AI to perpetuate biases, particularly regarding Indigenous peoples.
Protecting Creators’ Rights
As BAnQ moves forward, copyright concerns have surfaced as a significant issue for creators in the cultural sector. However, Grégoire has argued that the proposed platform could offer enhanced protection for artists compared to the current landscape, which she described as “the Wild West” in terms of data usage. The database is envisioned as a centralised gateway that could facilitate fair compensation for creators when their works are utilised in AI training.
Despite these assurances, some artists express unease about the potential implications for their livelihoods. Maxime Harvey, a postdoctoral researcher, highlighted a prevalent concern: “Even if artists earn income from it, they are still feeding the beast that will eventually replace contracts they may lose because of AI.” This sentiment underscores the delicate balance between innovation in AI and the preservation of artistic integrity.
Future Prospects and Funding
The timeline for the project’s operational launch is projected for 2029, although D’Amour indicated that this may be reassessed following the experimental phase. The estimated budget for the next five years is approximately $10.5 million, encompassing both operational and capital expenses. BAnQ has already secured $340,000 from the Quebec government for the feasibility study, along with an additional $750,000 to support the project’s initial experimental phase.
Why it Matters
This initiative is a pivotal step towards ensuring that Quebec’s rich cultural heritage is accurately represented in the rapidly evolving landscape of artificial intelligence. By building a dedicated database, BAnQ not only seeks to rectify the underrepresentation of Quebec society in AI systems but also to establish a framework that prioritises the rights of creators. As AI continues to shape our world, ensuring that it reflects the diversity and complexity of local cultures is essential for fostering inclusivity and understanding in the digital age.
