By Mark Buckshon
Ontario Construction News staff writer
OTTAWA – The artificial intelligence revolution in the architecture, engineering, and construction (AEC) industry may be inevitable, but it is colliding with a uniquely Canadian “trust deficit” and deep anxieties about the future of mentorship and professional liability.
That tension defined “AI in AEC: From Promise to Practice,” a packed extended-lunch conference hosted by the Ottawa Chapter of Construction Specifications Canada (CSC) at Algonquin College on Nov. 26.
While speakers urged the industry to adopt a “killer instinct” to solve Canada’s productivity crisis, the event’s most compelling moments came during a candid Q&A session that stripped away the hype to reveal the practical and human challenges facing the profession.
The efficiency paradox: Doing more, not less
Victoria Ikede, the National AI Lead at Architecture49, delivered a sobering reality check regarding the promise that AI will liberate professionals from drudgery.
Instead of freeing up time, Ikede warned of an emerging efficiency paradox where the technology simply raises the bar for output volume.
“I thought, ‘Oh, this is so much more productive, I’ll do less work,’” Ikede told the audience. “It’s the opposite, because now I feel like I can take on more work… I’m taking on more and more because I now have this assistant.”
She cautioned that while AI allows AEC professionals to produce impressive websites, renderings, or scripts instantly, it often robs them of the “struggle” required to truly master their craft.
“I’ve done all this work, but I don’t feel like I know anything,” she quoted a student as saying. “You lose the process that makes you confident as a person.”
The “Killer Instinct” and sovereignty
The event opened with a strategic call to action from Karen Savoie, president of Savoie Faire Consulting. She argued that Canadian firms, often culturally risk-averse, are falling behind global competitors who are aggressively integrating AI.
“There was a need for Canadians and Canadian businesses to adopt a ‘killer instinct,’ which kind of goes against our grain,” Savoie said, referencing themes from a recent University of Waterloo conference.
Emmanuel Florakas, a partner at BDO Digital, reinforced this with a technical roadmap. He advised firms to stop chasing flashy chatbots and start building a “single source of truth” for their data—an orchestration layer that cleans and organizes information so it can actually be used by AI tools.
Government regulation and initiatives
Jessica Wright, a senior policy advisor at Innovation, Science and Economic Development Canada (ISED), addressed the “trust deficit” head-on. Citing the 2025 Edelman Trust Barometer, she noted that only 30 per cent of Canadians currently trust AI.
Wright used the analogy of the automobile to explain the government’s role in the ecosystem.
“Cars get you from A to B faster… but in order to really benefit from that, you need roads,” she said. “And in order to have roads, you need standards, you need rules, you need to understand what the guardrails are.”
The mentorship crisis
The discussion took a critical turn when attendee Cristina Ureche-Trifu raised a fundamental question about the future of professional development.
“How do we teach people how to think?” she asked the panel, questioning how the industry can fill the learning gap created when AI automates the foundational tasks that junior staff traditionally used to learn the ropes.
Jeff Halashewski, a specification writer at Dialog and a CSC national vice-president, suggested that mentorship must evolve into a collaborative review process.
“It’s more or less in the mentorship, yes, going through that AI process, but taking it back to the beginnings of how we actually learn, just go line by line and validate,” Halashewski said.
The Infrastructure Challenge: AI’s Power Hunger
While much of the conversation focused on software, John Jensen, a CSC life member, raised a critical infrastructure concern that often goes overlooked: the physical power grid.
“It has been said that AI is going to require a lot of electrical power as it expands,” Jensen noted. “A concern is that our current infrastructure is not capable of providing the power that is going to be needed.”
In response, Florakas acknowledged the scale of the challenge, predicting a “re-excitement of nuclear energy” would be necessary to offset the massive computational demands of AI, alongside a general “rethinking” of the grid structure.
Lessons from history: “It takes longer than we think”
In a moment of historical perspective, John Cooke, a veteran structural engineer with John G. Cooke & Associates Ltd., urged patience. Correcting the room’s tendency to overestimate the speed of change, Cooke drew parallels to the industry’s sluggish transition to previous technologies.
“I look at the implementation of CAD… versus hand drawings in 1989 and the expectations of us at that time, and how long it actually took to get it implemented,” Cooke said.
He noted a similar pattern with Building Information Modeling (BIM). “We were all brought into a room by the government in 2013 and told we were going to produce this project in BIM… we’re still not there yet.”
“It takes longer than we think it does,” Cooke concluded, adding a practical concern about the massive electrical infrastructure required to support the exponential growth of AI computing.
The event was sponsored by CertainTeed and Ontario Construction News.