Quebec’s national library is embarking on an ambitious project aimed at creating a comprehensive database of cultural and governmental information. This initiative, spearheaded by Bibliothèque et Archives nationales du Québec (BAnQ), intends to improve artificial intelligence systems’ comprehension of Quebec’s unique society, culture, and Indigenous languages. Following a successful feasibility study, the library has entered the experimental phase of this endeavour, which could significantly address the shortcomings of existing AI technologies in representing Quebec accurately.
Aiming for Inclusivity in AI
The primary motivation behind this initiative is the recognition that many generative AI models often lack reliable data concerning Quebec. Valérie D’Amour, who directed the feasibility study, highlighted the necessity for a more inclusive approach: “All scenarios are a little bit on the table right now. We have a lot of ideas, and we want to validate the possibilities with cultural stakeholders, as well as with data owners and providers, who will be involved in the discussions.” This collaborative spirit aims to foster a rich dialogue among various cultural players, ensuring that the resulting data set is as comprehensive and representative as possible.
Marie Grégoire, BAnQ’s president and chief executive officer, reinforced the goal of cultivating AI systems that genuinely reflect Quebec’s societal values and cultural nuances. “That means having Quebec references, whether in small models or large models, whether they come from research or from the business community,” she articulated.
Learning from Global Initiatives
Similar projects have emerged globally, with notable examples in Sweden where vast collections of Nordic-language texts are being assembled to enhance AI models for Scandinavian languages. BAnQ plans to initiate the project with its own extensive collections before exploring additional data sources. This step is pivotal in creating a robust foundation upon which future AI models can be built, helping to ensure that they are both accurate and culturally sensitive.
The initiative springs from a 2024 report by Quebec’s innovation council, which pointed to the deficiency of available data on Quebec in the training datasets used by AI. Destiny Tchéhouali, co-holder of a research chair in French-language artificial intelligence and digital technologies, expressed concern about the current underrepresentation of Quebec culture in existing AI corpora. He warned, “We run the risk of reproducing linguistic biases and cultural biases. And when we also talk about Indigenous peoples, we run an even greater risk of all these biases.”
Addressing Copyright Concerns
As BAnQ forges ahead with its plans, questions surrounding copyright and the ethical use of artistic works have surfaced. Grégoire is optimistic about the potential for this new platform to provide better protections for creators. “Right now, it’s a bit like the Wild West. Data is being harvested for free, and that should not be the case,” she observed. By establishing a centralised database, BAnQ hopes to facilitate fair compensation for creators whose works are utilised in training AI systems.
Despite the potential benefits, some artists remain apprehensive. Maxime Harvey, a postdoctoral researcher at the National Institute of Scientific Research, noted that many fear their contributions could ultimately jeopardise their livelihoods. “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 said.
Looking Ahead
The feasibility study has set an ambitious target for the platform to become operational by 2029, although D’Amour remarked that the timeline would be reassessed post-experimental phase. The projected five-year budget stands at approximately £8.4 million (around $10.5 million Canadian), covering both operational and capital expenses. BAnQ has already secured £270,000 ($340,000) from the Quebec government for the feasibility study, along with an additional £600,000 ($750,000) to support the upcoming 12-month experimentation phase.
Why it Matters
This initiative represents a crucial step towards ensuring that Quebec’s rich cultural tapestry is accurately represented in the realm of artificial intelligence. By systematically addressing the gaps in data representation, BAnQ is not only safeguarding the province’s unique identity but also setting a precedent for how other regions can approach the integration of local content into AI systems. The potential impact on cultural preservation and the promotion of Indigenous languages cannot be overstated, as this project could serve as a vital blueprint for inclusivity in the rapidly evolving landscape of artificial intelligence.
