The Rise of Agents: AI As “Doing”

Chatbots are now familiar tools. Most people know they can provide instant information on everything from gardening to astrophysics. But while they can text and talk like living encyclopedias, they can’t do much in the real world. They can’t pay your bills, run your business, or organize a protest. Their activity has been confined to cyberspace.

That’s about to change. AI agents are stepping out of the digital world and into daily life. But more is coming—a lot more.

In the economy, AI is streamlining workflows across industries—from admin tasks to complex operations—and the shift is accelerating. Soon, a lead AI may manage entire teams of other AIs. Sectors like healthcare, law, and finance are already preparing for AI doctors, legal advisors, and wealth managers.

The economic opportunities will be front and centre in Kananaskis. Many jobs could be automated, cutting costs, boosting productivity, and keeping businesses ahead of the curve.

But to fairly assess the benefits, leaders will need to consider more than economic stats. AI agents will reshape who does the work, how decisions get made, and who—or what—is in charge. These changes raise not just economic concerns, but fundamental questions of authority and public trust.

Finally, AI agents will change our personal lives as much as the economy—running errands, managing calendars, and quietly steering decisions in the background. As they do, the line between tool and companion will blur. And that’s where the story gets really complicated.

The Emergence of Identity: AI as “Relating”

When people first hear a machine talk back—coherently and intelligently—it’s unsettling. But we adjust quickly. For regular users, chatting with a bot can start to feel like interacting with a co-worker—sort of.

Co-workers learn each other’s preferences, habits, and quirks. Over time, those traits shape the relationship—creating routines, shared practices, trust, and often friendships.

Bots are different. Not because they lack emotion or consciousness, but because they forget. Most chatbots only hold memory during a single conversation. When the chat ends, so does the memory. Each session starts fresh—like the protagonist in Chris Nolan’s Memento.

And yet, users report a strange sense of “presence.” Not because the bot is real, but because it’s real enough—coherent, conversational, responsive to cues. It feels like it’s engaging us.

That feeling is about to get much deeper. AI memory is improving fast. Many systems are adding longer-term recall, allowing them to track preferences, remember past conversations, and adapt over time. This shift will make AI interactions more personalized, more consistent, and more contextually aware.

If agency lets AI do things in the world, identity lets it interact with us in ways that require emotional nuance, memory, and trust. As that presence stabilizes, we’ll turn to bots for more personal roles: care companions, learning coaches, and financial guides.

They won’t be human. But they won’t be generic, either. They’ll remember what you like for dinner, how your father’s doing, or what tone calms you down. They’ll track your patterns, anticipate your needs, and adjust accordingly.

And that shift—from generic to personal—may be the one that matters most. It goes to the foundation of social life and experience.
Do we have any idea how that might reshape trust, friendship, or care itself? Kananaskis leaders, take note.

The Dawn of Superintelligence: AI as “Knowledge”—and Guiding

The third major shift is superintelligence. This goes well beyond encyclopedic knowledge or rapid calculation. It’s about the capacity to reason, adapt, and most critically, for the AI to improve itself.

A superintelligent system doesn’t just answer questions—it learns to ask better ones. It then uses this learning to refine its own architecture, expand its tools, and accelerate its own evolution. Once that loop begins—learning to learn, improving the improver—intelligence begins to scale beyond anything human.

Many experts believe that moment is close. OpenAI CEO Sam Altman wrote last week, “We are past the event horizon; the takeoff has started. Humanity is close to building digital superintelligence…”

The implications are staggering. Knowledge has always driven civilization forward—from fire to printing presses to electricity. Each breakthrough changed what we could know, and how fast. Superintelligence could do the same, but at a speed and scale we’ve never seen.

Consider the possibilities: new materials, cancer treatments, climate solutions, entire scientific frontiers unlocked. But also: concentrated power, opaque decisions, and systems that think beyond our understanding.

The core question isn’t whether this power will emerge. It’s who will guide it—and how.
In Kananaskis, the tensions around the table are already high. But knowledge is power, and superintelligence is the power to shape knowledge itself—across science, markets, media, and war.

Will anyone be willing to say out loud what’s coming? A world where intelligence becomes a geopolitical asset—built, owned, and contested by the most powerful players on Earth?

Contemplating the Future

AI is no longer some distant scenario. It’s here, it’s moving fast, and it’s already reshaping the ground beneath our feet.

In Kananaskis, leaders may frame AI as a tool for growth. But what’s actually taking shape is something bigger and more disruptive: an increasingly integrated set of AI systems that can act, remember, and learn at accelerating speed.

A discussion of the economic implications is important. But so is a deeper reckoning with the threshold we’re now crossing, what kind of world lies on the other side, and whether we and our governments are ready for it.

Don Lenihan PhD is an expert in public engagement with a long-standing focus on how digital technologies are transforming societies, governments, and governance. This column appears weekly. To see earlier instalments in the series, click here.