The underinvestment in and implementation of AI in the infrastructure sector.
Give a proud Italian any chance to loop Michelangelo into the conversation, and they’ll take it. But stay with me for a moment—I do have a point. If you hand me a hammer, a chisel, and a block of stone, my unskilled hands will deliver crushed rock. Hand the same tools to Michelangelo, and he will produce a masterpiece like the Statue of David. The unskilled worker has always blamed the tool and that is no different when it comes to artificial intelligence.
Blaming the Tool
When I hear colleagues in my industry express concerns about implementing artificial intelligence (AI), it often feels like they are missing the crucial element of the conversation. Much like blaming a hammer for not sculpting a masterpiece, too many focus on AI as the problem, rather than the training and expertise required to wield it effectively. There’s a pervasive misunderstanding that AI itself is the stumbling block when, in reality, the underinvestment and lack of proper implementation are what hold many industries back. Yes, I’m looking at you, infrastructure. While we are increasingly surrounded by technological innovations, Canada—despite being a global player in many industries—lags significantly behind in the adoption of AI. We’re still far behind sectors like aviation and manufacturing. Planes have had autopilot for the past 40 years, but similar automation is still missing in our cities’ largest infrastructure developments.
Let’s take our highways for example. Recently, a noise study was commissioned near a highway. To complete the noise survey, traffic was recorded for 24 hours, and the number of cars versus trucks was counted… manually. Despite the advanced technology and automation available today, a team had to interpret the data manually. Why?
The issue isn’t a lack of automated capabilities—it’s the challenge of properly integrating AI and machine learning algorithms to process the data effectively. The cameras and microphones provided valuable information, but without the right machine-vision algorithm, trained professionals to analyze the data, and—here’s the big one—budget to implement an upgraded system, the project relied on outdated methods. This is just one example of how technology has potential, but its success ultimately depends on budget, expertise, and a willingness to adapt.
The Challenge of Digital Transformation
About a year ago, when I posted a piece called Qualitative Schedule Risk Analysis vs. AI Schedule Risk Analysis in Major Projects, I stirred up a lot of controversy around the subject. I encourage you to go back and read some of the comment section and think about how far we’ve come in our thinking even in the last few months. While embracing the tools available to us now can feel threatening to those who don’t fully understand them, I am doubling down on embracing AI while knowing its limitations. This isn’t a “us versus AI” scenario—it’s a “us utilizing AI” scenario.
AI adoption in larger corporations is undoubtedly slow and I’ll admit, frustrating. Risk management plays such an important role. Are the resources and effort required to implement AI safely worth it? Does the cost of integration outweigh the risks of clinging to outdated methods?
AI gives us the ability to fail faster, learn faster, and ultimately improve faster. Every megaproject consists of hundreds of individual components that could, in theory, be automated—but it’s not that simple. Behind every AI implementation is a human. We are creatures of habit, and retraining individuals who have done things the same way for decades is no small feat.
Large corporations, in particular, often stifle innovation with bureaucratic barriers. This is why I’m deeply passionate about the agility of AI start-ups—their ability to move quickly, experiment freely, and challenge the status quo offers a glimpse into what’s possible when innovation is truly embraced. Corporations have more at stake, making hesitation understandable as they carefully assess and mitigate risk. But that hesitation—what I call “analysis paralysis”—is the enemy of progress. The CEO of OpenAI I didn’t wait for perfection before releasing GPT-3 or 3.5. He understood that to make AI widely accessible and usable, there had to be fewer barriers for adoption. This approach is key to how AI will continue to grow and evolve: embracing its imperfections and iterating over time.
The “Stone Soup” of AI Adoption
Have you ever heard of the Stone Soup folktale? In a recent episode of COMPLEXITY, the Santa Fe Institute Institute podcast, the hosts used the Stone Soup folktale to illustrate how intelligence and problem-solving emerge not from a single, fully formed idea but through collaboration and incremental contributions. This folktale, which dates back to at least the early 18th century, has appeared in various forms across Europe, with some of the earliest known versions originating in France and Portugal. The story tells of a weary traveler who arrives in a village with nothing but a cooking pot and a stone. Claiming to make a delicious “stone soup,” they place the stone in a pot of boiling water. The skeptical villagers watch as the traveler stirs the empty pot, eventually convincing them that the soup would be even better with just a few extra ingredients. One by one, the villagers contribute what little they have—carrots, onions, spices—until the pot transforms into a hearty, nourishing meal that none of them could have made alone.
This analogy perfectly mirrors the challenge of AI adoption and digital transformation today. Many organizations hesitate to integrate AI, not because the tools are inadequate, but because they lack the expertise, training, and collaborative mindset to make the most of them. Just as the villagers initially doubted the traveler, businesses often see AI as an uncertain investment rather than an opportunity to build something greater. The middle-aged professional may view the technology as a threat rather than an added advantage to what they’re already doing. The unskilled worker may blame the tool when results fall short, but the real issue lies in how it’s used. Without the villagers’ ingredients, there is no soup—just as without the expertise of professionals in infrastructure, artificial intelligence holds no real value.
So, now what?
Tools, whether it’s a hammer, a chisel, or an AI algorithm, are only as effective as the skill of the person using them. This is where the importance of proper training comes in. When we talk about AI, we must remember that it’s not just the tool but the expertise in using it that makes a difference. It’s easy to criticize AI as a flawed or ineffective tool, but it’s the lack of proper integration, training, and a willingness to innovate that’s often the real barrier.
As I continue my studies in AI at Oxford, our professor often asks, “What changed in AI since last night?” This question reflects the rapid pace of AI development, but it also highlights the cyclical nature of technological advancement. AI has been around for decades, going through phases of intense excitement and long periods of stagnation. We’re currently in a phase of “summer” for AI, but who knows how long this will last? The key is to start embracing AI now, with a recognition of its limitations and the willingness to continually learn.
Don’t Blame the Hammer
In the end, AI is like any tool—it requires the right hands to wield it effectively. We can’t simply blame the tool for not delivering results if we haven’t invested in the proper training or set the right expectations. If we want to see real transformation in our industries, it’s crucial to move beyond skepticism and embrace the potential of AI, with the understanding that, like Michelangelo, it’s not the tool but the expertise behind it that truly creates masterpieces.
The future of AI will be built not on the stone itself, but on the collective contributions of those who understand its power and limitations. It’s time we stopped blaming the tool and started learning how to use it properly. Only then can we unlock its true potential.