Every technological leap in history has come with promises and social tensions, and the digital transformation now underway is no exception. What makes today’s technology different, however, is the speed of change and the concentration of ownership by a handful of corporations, accelerating the inequalities that already existed in our societies. This concentration is happening while we experience a broader crisis in the neoliberal economic order, where control over knowledge itself, rather than traditional production, has become a central axis of power.
The scale of investment now flowing into artificial intelligence and the “digital stack,” the layered infrastructure underpinning these tools, has taken on the character of an arms race, with U.S. and Chinese firms competing for global dominance. Meanwhile, automation and the growth of gig-economy labour relations continue to erode job security, extending the sense of precarity into sectors once considered insulated from disruption, including highly skilled professions. This is compounded by the weakening of the state itself through layoffs of public employees, as is currently occurring in Canada, where more than 40,000 workers are expected to lose their jobs, carried out under the inference that AI tools can take over their functions.
Online platforms compound the challenge: what began as tools for communication have become the very infrastructure of public debate, shaped by algorithms that exploit human negativity bias to maximize engagement—structurally amplifying divisive and emotionally charged content not by deliberate design, but as the predictable output of optimizing for attention. Disinformation is not an anomaly of this system; it is its product. The corporations leading the AI race are frequently the same ones that control these platforms, granting them unprecedented influence over which voices are amplified and which version of reality becomes dominant, with direct consequences for electoral integrity and institutional stability.
Faced with this, national governments have generally responded with limited and reactive measures, most confined to personal-data protections that fail to address the deeper structural imbalance. It is precisely this gap that the Broadbent Institute and Rumbo Colectivo set out to address last week in Toronto, organizing the seminar “Democratic AI & Digital Sovereignty” as the third preparatory meeting on the road to the 2026 Panamerican Congress. The event brought together legislators and experts from Canada, Chile, Costa Rica, and the United States, including local and federal representatives from Canada, to begin building shared guidelines on democratic artificial intelligence and digital sovereignty.
The initiative reflects an urgent correction of course. Currently each country is confronting – inefficiently – these dynamics in isolation. In contrast to this, the workshop sought to build a hemispheric response capable of matching the coordinated, global scale at which major tech corporations already operate.
The guidelines developed through this Toronto meeting are not an endpoint but a building block. They are intended to move directly into legislative and inter-parliamentary debates and, ultimately, into the platforms of progressive governments across the continent, addressing central questions about whether we are willing to challenge private capital financing technological innovation. Only this approach will be able to prevent automation from deepening labour precarity without rejecting innovation outright, and shift the political question from what to moderate to who owns and governs the infrastructure of public debate.
This work now points directly toward Montevideo, where the Panamerican Congress will convene delegations from fifteen countries between August 21 and 23, under the leadership of President Yamandú Orsi. There, the guidelines shaped in Toronto will be presented to a broader hemispheric audience, marking a further step toward a coordinated response—one that must go beyond shared regulation to include shared infrastructure: public platforms built on open protocols, with recommendation systems designed to bridge rather than divide, governed multilaterally rather than by private boards.
