The Digital Governance Standards Institute (DGSI) has announced the release of the first Canadian standard of its kind addressing the operational and research uses of artificial intelligence (AI) and machine learning (ML) technologies in public-serving research institutions.
The new national standard – “CAN/DGSI 128:2025 – Machine Learning and AI Implementation in Research Institutions” – aims to provide foundational guidelines for the ethical, transparent, and sustainable use of AI and ML technologies in research environments such as libraries, archives, universities, and laboratories, according to DGSI’s news release.
“This standard will help to establish initial guidelines for research institutions, especially smaller institutions, to begin safely and effectively navigating AI adoption in trustworthy and human-centred ways,” said Matthew da Mota, chair of the expert drafting team and senior research associate of the Centre for International Governance Innovation (CIGI), in the news release.
The new standard covers the following:
- the minimum requirements for organizations utilizing AI tools to manage, preserve, analyze, or distribute research data and content
- risks like information loss, algorithmic bias, research security, and environmental effects
- the necessity of promoting transparency, oversight, and cultural and data sovereignty
- operational uses of AI systems, including cataloging, digitization, and research infrastructure
- research uses of these systems, such as data analysis, machine-generated insights, and digital experimentation
DGSI said the standard’s sections seek to help organizations:
- Preserve information and data integrity in physical and digital environments alike
- Keep metadata, provenance, and historical context across shifting research systems
- Enable informed, equitable, and rights-respecting access to AI-enhanced tools and services
- Impose sustainability requirements for AI system selection and operation
- Implement user feedback mechanisms and other oversight, audit, and transparency practices
- Offer training and literacy to drive responsible experimentation and cross-disciplinary collaboration
Development
In its news release, DGSI shared that it convened the national expert committee that developed the standard, which aims to address the need for responsible AI governance in sectors managing knowledge, facilitating discovery, and acting as public digital infrastructure.
“Research institutions are stewards of collective knowledge and critical infrastructure in the digital age,” said Darryl Kingston, DGSI executive director, in the news release. “This standard supports their mission by helping them adopt AI tools in ways that protect research integrity, institutional autonomy, and public trust.”
”As Canada plans to build a more sovereign and secure technology ecosystem, we need to pay close attention to how we support and protect our world-class research institutions, and the information and data they house,” da Mota added.
In its news release, DGSI noted that the new standard shows how Canadian research institutions are uniquely managing extensive data amounts and historical materials while making advancements in public interest science. DGSI added that the standard complements its current frameworks on ethical AI and responsible data stewardship, among others.