AI Superintelligence Is Coming—and Canada Is Super-Unready.

  • National Newswatch

Thanks to Mark Carney, AI is finally on Canada’s radar—which is good news. His new Minister of AI, Evan Solomon, says we’re at a “Gutenberg Moment,” and promises to transform the economy through sovereign cloud and domestic data centres. It’s a worthy goal. But AI sets its own agenda—which can shift in a heartbeat. Just ask Mark Zuckerberg.

After his new chatbot (Llama 4) performed poorly this spring, the Meta CEO appeared to hit a wall—then pivoted sharply: the AI race, he realized, is moving beyond chatbots. Suddenly, it’s about superintelligence (SI), and any company that’s not in that race will be left behind. 

Artificial Superintelligence refers to an AI system that surpasses human intelligence across all domains—creativity, problem-solving, reasoning, and even emotional intelligence. And Zuckerberg is now moving like a whirlwind to align Meta’s priorities with that vision. 

In the last few weeks, he’s torn up his strategy, sidelined senior leaders, and committed $12 billion to an SI moonshot—to start. He’s been hiring aggressively, offering signing bonuses of up to a $100 million. He even tried to buy Safe Superintelligence outright for $34 billion. Its co-founder, Ilya Sutskever—one of the most respected minds in the field—turned him down. Which raises the question: What must you believe you’re building to say no to $34 billion?

That’s the deeper question now echoing through Silicon Valley: What kind of intelligence are we building—and who will control it?

The biggest names in AI—Google, OpenAI, Microsoft, xAI, Anthropic—are all chasing the same prize. Not just smarter tools. Not even Artificial General Intelligence. The target now is something else entirely: intelligence that improves itself—and that’s a quantum leap. 

If Canada wants a role in shaping this future, first we need to understand what kind of system is being built.

What Is Superintelligence?

SI isn’t a better search engine or a clever chatbot. It’s what happens when intelligence breaks free of human constraint and begins to improve itself—when systems can learn, reason, and evolve beyond anything we can build, or even fully understand.

If the industrial revolution replaced muscle, SI replaces thought. It asks better questions, finds patterns no human can see, and teaches itself. And that sets it apart from every other kind of AI in two basic ways.

First, it can improve itself. Today’s models depend on human retraining. Superintelligence won’t. It will rewrite its own architecture, refine its logic, and keep getting smarter—on its own.

Second, it generalizes. Current AI is narrow: good at code, or language, or images. Superintelligence connects across domains—science, policy, biology, engineering—solving problems no single expert could tackle.

The combined result is an accelerating spiral of learning and improvement—with no known ceiling. Imagine a system that can:

  • Devise new scientific theories
  • Simulate entire economies or ecosystems
  • Invent medicines, materials, or models of government
  • Model global tradeoffs in real time
  • Redesign itself to do it all better tomorrow

Zuckerberg’s plan reflects this vision. His new Meta Superintelligence Labs (MSL) isn’t looking to fine-tune his chatbots. It wants to build the machine that will build all other machines.

Some say we’ll reach this level within five years. Others say it’s already arriving. But among the people building these systems, a consensus is forming—and the timelines are tightening. Here’s what some of the top CEOs are now saying:

  • Open AI CEO Sam Altman recently wrote that superintelligence will arrive “in a few thousand days.” In June, he added: “We have recently built systems that are smarter than people in many ways... We are about to find out how far beyond human-level intelligence we can go.”

     
  • Anthropic CEO Dario Amodei says AI may soon function like “a country of geniuses,” capable of nearly every human task. He projects these capabilities will emerge “within the next two to four years.”

     

  • xAI CEO Elon Musk states that “AGI is probably next year, within two years... Probability that AI exceeds the intelligence of all humans combined by 2030 is ~100%.”

 

 

  • Safe Superintelligence CEO Ilya Sutskever declares that: “AI will be extremely unpredictable and unimaginable.”

As for Canada, while the Carney government has committed to AI, SI barely figures in the national conversation. It’s fair to ask if there’s even a place for us in this race.

Why It Matters

The question is not idle. SI isn’t just a breakthrough. It will create a new kind of geopolitical leverage—a paradigm shift that will transform more than knowledge: it will transform power relationships

SI is not just good at solving hard problems. It knows which problems matter—and that changes how we use knowledge. SI doesn’t just find answers, it shapes agendas. It defines the conversation, sets the objectives, and directs the game. It controls the board.

This, in turn, means it can out-strategize generals, out-negotiate diplomats, out-diagnose doctors, and out-code engineering teams. SI sees further. It adapts faster. It learns across domains.

Moreover, an SI system keeps improving itself—week by week—with no ceiling we can yet define. Whoever builds SI first will gain extraordinary capacity—not just to solve problems, but to coordinate systems, shape institutions, and outmaneuver rivals. They’ll have leverage over everything that runs on knowledge: science, defense, finance, diplomacy, energy, media, governance.

That’s the vision now driving many of the world’s top AI labs: not just intelligence that is supersmart, but intelligence that is super-strategic. That’s why Meta created MSL—and why Google DeepMind, OpenAI, xAI, and Anthropic are doing the same. 

And Canada?

What Canada Can Do

While we’ve graduated to cloud infrastructure and digital regulation, superintelligence is barely on the agenda. It should be: this isn’t just about keeping up; it’s about who will define the future. 

Canada helped launch modern AI. Researchers like Geoffrey Hinton and Yoshua Bengio were global pioneers. Institutions like MILA and the Vector Institute remain respected leaders in ethical AI.

But when it comes to SI, we’re all but absent. No readiness strategy. No safety framework. No international convening effort. And little sense of urgency.

This isn’t just a policy lag. It’s a visibility gap. A strategic one. And it’s growing.
We don’t need to match Big Tech dollar for dollar. But we do need to get in the room where the real conversation is happening—and help shape it. There are cards we can play. Here are five credible moves Canada could take right now:

  • Plan for What’s Coming
    Create a national strategy to prepare for superintelligence: train experts, forecast risks, and build the tools we’ll need to respond.

     

  • Test These Systems for Safety
    Set up an independent institute to stress-test powerful AI models—working with universities and international partners to figure out what’s safe, and what’s not.

     

  • Build Public AI Infrastructure
    Invest in computing power and shared tools that serve the public—not just tech giants. Help Canadian researchers stay in the game.

     

  • Ensure AI Transparency and Accountability
    Make sure powerful AI systems are subject to clear rules, public oversight, and independent audits—so Canadians know how decisions are being made, and by whom.

     

  • Use Our Voice Globally
    Canada has convening power. Let’s use it—by helping lead international efforts to set rules, share knowledge, and keep AI aligned with public values.

If the experts are to be believed, and if money—very big money—really talks, SI is coming—soon. The question now is not just who builds it, but who helps shape how it gets used. There is still a place for Canada—if we choose to claim 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.