Quebec's national library is moving ahead with plans to create a database of cultural and government content aimed at training AI systems to better understand Quebec society, culture, and Indigenous languages.
Bibliothèque et Archives nationales du Québec (BAnQ) has launched the experimental phase of its proposed government and cultural databank. The project addresses concerns that major generative AI systems often fail to provide reliable information about Quebec due to limited data.
"All scenarios are a little bit on the table right now," said Valérie D'Amour, who led the feasibility study. The platform would not serve as a public distribution channel; access would be tightly controlled. Marie Grégoire, BAnQ's CEO, said the goal is to ensure AI systems better reflect Quebec society.
The initiative stems from a 2024 report by Quebec's innovation council, which cited the "very small quantity of data on Quebec" in AI training datasets. Destiny Tchéhouali, a professor at Université du Québec à Montréal, warned of risks of linguistic and cultural biases, especially regarding Indigenous peoples.
Copyright concerns persist. Grégoire argued the platform could offer creators greater protection than the current "Wild West" environment, where data is harvested for free. The feasibility study estimates a five-year budget of nearly $10.5 million, with the platform targeted to become operational by 2029.