“AI progress will be immense from here…”

Sam Altma

Sam Altman, CEO of OpenAI, claims we’re on the brink of a breakthrough he calls “Reasoning AI” (also click here). If true, this could be historic, posing a crucial question for Canada's next election: Do Canadians welcome or fear the rise of Reasoning AI? Let’s start with some context.

The End of AI Chapter 1 

When ChatGPT was released in November 2022, it stunned the world by rallying 100 million users in three months, making it the fastest-growing app in history. But the excitement quickly turned to fear as concerns about rogue AI, job losses, and disinformation spread.

In March 2023, an "Open Letter" signed by hundreds of AI experts called for a global pause on AI research, citing risks to humanity. The public was quickly divided into "doomers" and optimists, cutting across party and ideological lines. For instance, the threat to jobs might seem anti-union, but AI is more likely to impact non-unionized professionals like lawyers, accountants, and programmers. Similarly, while some see AI as pro-business due to productivity gains, many progressives are drawn to its potential to revolutionize public services like healthcare and education.

In short, people's views on AI cut across social, economic, and educational lines, often dividing families and friends rather than aligning with traditional political affiliations. With rogue AI, these profiles are even harder to pin down, as the issue provokes intense emotions for complex and varied reasons. These attitudes may be more dispositional than ideological.

In the past six months, the hype has receded, and doomsday scenarios have quieted, but the tensions between skeptics and optimists remain unresolved. According to Altman, we’re at a turning point— the end of AI's “first chapter” and the beginning of a new one. Going forward, he promises, progress will be “immense.” Let’s briefly consider why, before turning to the political implications. 

The Dawn of AI Chapter 2

Chatbots are trained on vast amounts of data stored in their memory, generating responses based on billions of connections. For example, current AI can sift through extensive medical datasets related to cancer and recognize patterns. However, its capacity for analysis stops there—it cannot combine these patterns in new ways to find a cure. This kind of problem-solving requires reasoning.

Reasoning AI opens a new chapter by merging pattern recognition with the ability to plan and solve problems. According to OpenAI, achieving full-fledged human-like reasoning (Artificial General Intelligence) involves five levels. We are currently at Level 1, but on the brink of Level 2, which will enable AI to solve problems at a PhD-level human intelligence.

As AI advances from Levels 2 to 5, it will begin setting its own goals, making plans, and acting on them. By Level 5, AI will be capable of executing highly complex tasks, such as running a corporation or large research project, and even improving itself.

The Election and AI’s New Frontier

From a political perspective, the implication is clear: Reasoning AI represents a quantum leap toward autonomous AI—one almost certain to reignite fears about rogue AI, job losses, and disinformation. A machine with such advanced capabilities poses a far greater threat than chatbots, raising the stakes for AI opponents.

But the optimists' scenario is equally compelling. Full-fledged Reasoning AI could transform our world, offering new tools to manage our increasingly interdependent systems. Traditional methods can no longer cope with the complexities of globalization; AI has the potential to modernize and manage our institutions effectively.

In this view, Reasoning AI heralds hope for a new era, enabling groundbreaking advancements in scientific research, supply chain management, environmental sustainability, and healthcare reform. It is the crucial instrument needed to address these global challenges. Without such advanced AI, globalization and democracy may falter, possibly reverting to more authoritarian forms of governance. Last week's column on globalization explored these issues and the challenges they present.

So where are Canadian political leaders likely to stand?

The world’s future is now too tightly tied to AI to reject it. Abandoning AI would mean falling behind in global progress and innovation. Given the significant investments and rapid development in AI worldwide, even a pause seems improbable. Governments will likely seek a "balanced" approach, ensuring safety while fostering advancement.

For Canada, finding this balance poses a special challenge. OpenAI’s CTO, Mira Murati, reports that Level 2 AI is only 12-18 months away, potentially aligning its release with the upcoming election. In short, Canadians may already be hurtling toward a do-or-die debate between AI skeptics and optimists.

Furthermore, many experts, including Altman, believe Levels 3, 4, and possibly 5 could be achieved by the end of the decade—in other words, they will emerge during the next government’s term. Political parties' campaigns must therefore go beyond the immediate impacts of Level 2 AI and seek a mandate to respond to the emergence of full-fledged Reasoning AI.

It is unclear what such a mandate might include, but it will be necessary despite the prospect of fractious debates and further public division. Political parties will struggle to take clear positions as traditional ideological lines blur. Some Canadians will see full-fledged Reasoning AI as their worst nightmare; for others, it promises a new and prosperous future. Underestimating the emotional commitment on either side would be a mistake.

Of course, the future remains uncertain. If Altman is to be believed, the AI clock is ticking, and we are fast approaching a turning point. There is an old saying in politics that hope always wins out over fear. Soon, our politicians may be confronted with hope and fear in their starkest forms. How will they respond? For now, we watch and wait.

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.