Ekaterina: Your book Latin America’s Left Turns describes how economic discontent fueled populist movements. Could AI-driven automation produce similar political backlash?

Prof. Cameron: It’s very possible.

For decades, the promise made to blue-collar workers was that globalization and technological change would create new opportunities. In reality, many workers who lost manufacturing jobs found themselves with lower wages, worse employment conditions, or outside the workforce entirely. That experience generated deep resentment toward political and economic elites.

What we may be seeing now is something similar affecting white-collar professions. Automation has the potential to displace not only factory workers but also coders, analysts, and many other professionals.

If that happens without strong support systems - retraining, relocation assistance, social protections - it could generate a comparable backlash. The political consequences could be profound.

Ekaterina: North American countries tend to produce AI technologies, while many countries in Latin America primarily import them. Does that resemble earlier patterns of economic dependency?

Prof. Cameron: In some ways, yes.

A common comparison is the mobile phone revolution. In countries like Peru, getting a landline installed used to take years. When mobile phones arrived, they allowed societies to leapfrog an entire stage of infrastructure development.

Something similar could happen with AI. In principle, it could give countries powerful tools without requiring decades of technological buildup.

But the question is always: who benefits, and at what cost? The distributive consequences matter. Who can afford these technologies? Who controls them? And what environmental or social costs accompany them?

Another concern is linguistic and cultural bias. Much of the data used to train AI systems comes from English-language sources. That can subtly shape the kinds of knowledge and policy solutions the systems generate.

Ekaterina: Does that bias appear within countries as well - for example, in debates about AI in policing?

Prof. Cameron: I saw an example of this in a classroom exercise. In a course on Latin American politics, students conducted a simulation of negotiations within the Organization of American States about organized crime.

I encouraged them to use AI tools to help draft policy proposals. What emerged was interesting: many of the policy papers ended up looking quite similar. They tended to emphasize surveillance technologies, enforcement mechanisms, and border controls.

What was largely missing were discussions of the underlying social and economic drivers of organized crime - poverty, displacement, agricultural systems, migration patterns.

That suggests the training data behind these tools may subtly steer thinking toward certain policy frameworks. Students could push back against that, but it required conscious effort.

Ekaterina: If we project twenty years into the future, do you think AI will deepen democratic inequalities between countries - or affect democracies more evenly?

Prof. Cameron: My expectation is that the effects will be uneven.

In theory, AI could enable more responsive governance. If a system could accurately capture the preferences of millions of citizens across a wide range of policies, it could generate policy recommendations that reflect those preferences more precisely than elections currently do.

That’s the optimistic scenario.

The pessimistic one is that AI becomes controlled by a small group of extremely wealthy actors who use it to advance their own interests - effectively creating a new form of oligarchic power.

Where countries fall along that spectrum will depend less on the technology itself and more on political institutions: regulatory frameworks, education systems, and the strength of democratic norms.

Countries with robust democratic governance are more likely to harness AI as a positive force. Where institutions are weaker, the risks are significantly greater.