AI, the Human Brain, and the Future of Open Science: An Interview with Prof. Guillaume Dumas, Researcher in Computational Psychiatry and Social Neuro-AI
What if computers could help us understand the very organ that made them possible: the human brain?
Dr. Guillaume Dumas, a researcher at the intersection of neuroscience, artificial intelligence, and collective intelligence, believes we’re just beginning to uncover what advanced computing can do for brain science, and what brain science can teach us about AI. He is Associate Professor in Computational Psychiatry of the Faculty of Medicine at the University of Montreal and the Director of the Precision Psychiatry and Social Physiology laboratory in the CHU Sainte-Justine Azrieli research center.
The Brain as a Blueprint for AI
Dr. Dumas explains that his research uses neural-inspired architecture, applying the principles of how the brain works to design smarter AI systems.
“Advanced computing has been a key tool for simulating both realistic brain models and abstract cognitive architectures,” he says. “As AI depends on computing, neuro-inspired AI is no exception.”
By using high-performance computing (HPC) to simulate how neurons communicate, researchers like Dumas can test hypotheses about learning, memory, and social behavior — insights that may help design more cooperative, human-like AI systems.
The Power of Many Minds — and Many Brains
A central concept in Dumas’s work is collective intelligence — the idea that groups can think and solve problems better than individuals.
“It’s like a team brainstorming session,” he explains. “The diversity of minds, skills, and perspectives increases the efficiency of collective problem-solving.”
To study this, Dumas developed a technique called hyperscanning, which records brain activity from multiple people simultaneously as they interact. Several different social tasks recorded include imitation of hand movements, sharing emotional anecdotes, mother-baby interactions, verbal communication, and non-verbal communication. His research showed for the first time that the brains of people synchronize during social interaction, a discovery that reshaped how scientists think about human connection.
“When we demonstrated the emergence of inter-brain synchrony during social interaction in 2010, many people didn’t believe it,” he recalls. “Now our study is one of the most cited in the field.”
Computing Inequality in Science
While breakthroughs like these rely on powerful computing tools, Dumas warns that access remains deeply unequal.
“If you don’t have access to high-performance computing, you can’t work as fast as well-funded labs or industry,” he says. “That’s a major challenge for early-career researchers.”
Open Science and the Public Good
For Dumas, the solution lies in open science, a principle he has championed for years.
“To me, ‘open science’ is almost redundant. Science should be open by nature,” he says. “Data sharing is critical for all research, including brain science.”
He believes that the fruits of research should be open-source and publicly accessible, so that discoveries benefit everyone, not just large corporations or wealthy institutions.
That’s precisely the vision behind Computing for Humanity: to make advanced computing a public good, available to anyone advancing science for humanity’s benefit.
Ethics, Ecology, and Education
When asked about ethical concerns in computational neuroscience, Dumas offers a nuanced perspective.
“We’re far from having the kind of ethical crises you see in industrial AI,” he explains. “The main issues right now are data privacy and ecological impact — making sure our simulations don’t waste energy.”
He notes that while the public often fears “artificial consciousness,” such fears are premature.
“It’s not the models that are the problem; it’s more that people tend to attribute spontaneously consciousness to things that behave like a human. On top, tech companies are playing on this. And now many people are getting attached to AI chatbots. But to me it's more like a human problem than a model problem. “ He says. “Education and critical thinking are key.”
Looking Ahead: Computing as a Public Resource
When asked what could accelerate neuroscience the most, Dumas is clear:
“If computing were treated as a universal public resource, we’d see faster discovery of biomarkers, better disease modeling, and stronger collaboration across disciplines.”
He envisions a future where computing, storage, and expertise are part of the same equation, a shared foundation for scientific progress.
Why This Matters
At Computing for Humanity, we believe that advanced computing should not be a privilege, it should be a shared resource that drives innovation for all. Researchers like Dr. Guillaume Dumas remind us that understanding the brain, improving health, and advancing AI are interconnected goals — and equitable access to computing is the key to achieving them.
This is where nonprofit-driven infrastructure becomes essential. Platforms like myresearchcloud.ca, launched by Computing for Humanity, aim to level the playing field by providing affordable and accessible computing resources for researchers and citizen scientists who lack institutional support.