← Toutes les nouvelles
April 7 – World Health Day
Building Trust: Where Does Responsible AI in Healthcare Stand, 7 Years Later?
On World Health Day, we revisit the key insights from a conference presented at MTL connect 2025, at the intersection of ethical, governance, and digital sovereignty issues.
April 7 marks World Health Day. An opportunity to revisit the discussions from the conference « Building Trust: Responsible AI in Healthcare, 7 Years After the Montreal Declaration ».
Moderated by Gaëlle Hermans, a pedagogical designer at CHUM, Catherine Régis, a professor at the Université de Montréal, and Jean-Louis Fraysse, founder of Bot Design, this conference provided a clear overview of the evolution of artificial intelligence in healthcare.
In 2017, Montreal laid the groundwork for global reflection with the Montreal Declaration, a pioneering ethical compass to guide the development of artificial intelligence. Seven years later, the landscape has profoundly changed. AI is no longer a prospective concept; it is now integrated into clinical practices.
From Ethics to Sovereignty
Drawing on their expertise in governance and AI system design, Catherine Régis and Jean-Louis Fraysse describe this transformation in three phases:
- Ethics Season (2017–2022): emergence of guiding principles
- Regulation Season (2022–2025): legal structuring (AI Act, Canadian frameworks)
- Adoption Season (today): concrete integration into healthcare systems
This evolution now raises a central question: how can we ensure responsible adoption without compromising data and medical decision sovereignty?
The Canadian Paradox
As the Speakers highlighted, according to a KPMG report (2025), Canadians want appropriate regulation to increase their trust in AI, yet the country ranks between 25th and 27th out of 30 OECD countries in terms of AI adoption.
A structuring paradox emerges:
Regulation builds trust, but it can also slow adoption and limit technological autonomy.
AI Already Present in Clinical Practice
The observations shared by Catherine Régis and Jean-Louis Fraysse are unequivocal: nearly 50% of medical and pharmacy interns already use tools like ChatGPT in their practice.
The question is no longer “should we use AI?” but: how can we integrate it reliably, ethically, and sovereignly into healthcare systems?
Very Real Risks
The two experts also highlighted several major challenges:
- Algorithmic bias: under-diagnosis in certain populations, particularly in dermatology
- Technological instability: rapid evolution of models, difficult to regulate
- Cognitive dependence: increasing delegation of clinical judgment
Added to this is a strategic issue raised during the discussion:
dependence on external technologies, which raises the question of control over health data.
Toward Sovereign Digital Healthcare
Beyond technical issues, the conference highlights a fundamental question:
Who controls the data, tools, and decisions? Building responsible AI also means strengthening the collective ability to master digital infrastructures, protect sensitive data, and preserve decision-making autonomy in healthcare systems.
Extending the Reflection at MTL connect 2026
These issues of trust, governance, and digital sovereignty will be at the heart of discussions at the next edition of MTL connect.
To go further and actively participate in reflections on digital sovereignty in healthcare, join us at MTL connect 2026. From October 13 to 16, 2026, in Montreal