Quebec’s Bibliothèque et Archives nationales du Québec (BAnQ) is embarking on an ambitious project to develop a comprehensive database that will include cultural and governmental content aimed at improving artificial intelligence systems’ comprehension of Quebec society, culture, and Indigenous languages. This initiative, which has now entered its experimental phase after a successful feasibility study earlier this year, seeks to remedy the inadequacies in AI training datasets that often overlook the rich tapestry of Quebec’s unique identity.
Addressing the Data Deficiency
The move comes amid growing concerns regarding the limited representation of Quebec-related data within major AI frameworks. Valérie D’Amour, who spearheaded the feasibility study, expressed optimism about the project’s potential. “All scenarios are a little bit on the table right now,” she noted, highlighting the importance of engaging with cultural stakeholders, data owners, and providers to validate the various possibilities.
The proposed databank is designed not to serve as a public distribution platform for creative works; rather, access will be tightly regulated to ensure the integrity and security of the information. Marie Grégoire, BAnQ’s president and CEO, underlined that the initiative aims to ensure AI systems accurately reflect Quebec’s societal nuances. “That means having Quebec references, whether in small models or large models, whether they come from research or from the business community,” she stated.
Learning from Global Examples
BAnQ’s initiative is part of a broader trend, mirroring similar efforts seen in other regions, such as Sweden, where extensive collections of Nordic-language texts are being compiled to support the development of generative AI models. Initially, BAnQ plans to utilise its own collections before expanding to include external data sources.
The drive for this project originates from a recommendation made in a 2024 report by Quebec’s innovation council, which acknowledged the scarcity of Quebec-specific data in AI training datasets as a major obstacle. Destiny Tchéhouali, a co-holder of a research chair focused on French-language artificial intelligence at Université du Québec à Montréal, emphasised the risks of linguistic and cultural biases prevalent in current AI corpora. “Quebec culture remains underrepresented in the corpora currently circulating in the AI world,” he remarked, underlining the urgency for this database.
Navigating Copyright and Creator Concerns
As BAnQ moves forward, it faces pressing questions about copyright and the implications for artists and creators. Grégoire contended that the proposed platform could enhance protections for creators, countering the current “Wild West” environment in which data is often used without compensation. “Data is being harvested for free, and that should not be the case,” she argued, suggesting that a centralised gateway could facilitate fair remuneration for creators whose works contribute to the database.
However, some artists express unease about the potential ramifications of contributing their work to AI systems. Maxime Harvey, a postdoctoral researcher, voiced the prevailing concern that artists might inadvertently undermine their own livelihoods by feeding AI models that could eventually replace traditional contracts. “The main criticism we hear in the field is that, even if artists earn income from it, they are still feeding the beast that will eventually be used to replace contracts they may lose because of AI,” he cautioned.
The Path Ahead
The feasibility study anticipates that the platform could be operational by 2029, although D’Amour indicated that this timeline would be reassessed following the experimental phase. The project has been allocated a five-year budget of approximately £8.5 million (C$10.5 million) through 2030, which covers both operational and capital expenses. Financial backing includes £260,000 (C$340,000) from the Quebec government for the feasibility study and an additional £600,000 (C$750,000) to support the initial 12-month experimentation phase.
Why it Matters
This initiative represents a crucial step towards ensuring that Quebec’s diverse cultural landscape is accurately reflected in the rapidly evolving field of artificial intelligence. By addressing the data shortages and biases that have long plagued AI systems, BAnQ is not only safeguarding the province’s cultural heritage but also paving the way for a more equitable and representative digital future. In a global context, this endeavour could serve as a model for other regions striving to preserve their unique identities in an increasingly homogenised digital world.
