If AI (or especially the hypothetical human-level AI or “AGI” in the future) becomes the new ‘digital oil’, it will function as a new ‘extractive resource’
What are the chances of our democracy surviving another 20 years? There is a rule of thumb, based on a century of history, that helps us measure this risk. And the future of AI has everything to do with this.
A study by the Journal of Democracy shows that democratic stability does not depend on “gross” GDP, but on “productive” GDP. In other words, wealth that comes from real economic activity (industry, services, innovation), and not from the simple extraction of concentrated resources, such as oil and gas or ores.
The rule is clear: when this “productive” GDP exceeds US$20,000 per person, democracy has about a 97% chance of surviving another 20 years. Below that, the risk increases greatly. The lesson is that diversified economies, where wealth circulates, create a much larger “institutional cushion”.
Where is Brazil? On a knife’s edge.
Our “productive” GDP, already discounting our heavy dependence on commodities, is around US$20,770. We crossed the line, but barely. We are technically in the safe zone (around 97%), but without any clearance for shocks.
This is where Artificial Intelligence comes in.
If AI (or especially the hypothetical human-level AI or “AGI” in the future) becomes the new “digital oil” — extraordinary but hyper-concentrated wealth in the hands of a few platforms — it will function as a new “extractive resource.” It sucks income from the rest of the economy and does not strengthen institutions, in other words, it would become a new “natural resource curse”.
In a high concentration scenario, our real “productive” GDP could fall to levels seen today in much more fragile democracies, reducing our chances of democratic stability from 97% to close to 75% or even less. Technological progress, if poorly managed, can erode our democratic cushion.
On the other hand, we can avoid the “AI curse” and turn it into a blessing. The goal should be to use AI to increase everyone’s productivity, not the income of a few.
This requires policies aligned with economic freedom: ensuring competition and interoperability, with attentive antitrust; use state purchasing power to foster open infrastructure; and focus on disseminating technology to small and medium-sized companies (via BNDES, Sebrae and SENAI), and not just the “national champions”.
An eventual Universal Basic Income policy may also become necessary in the future, especially for those who have their jobs automated by AI, in addition to well-designed legislation that protects the population from its already existing negative effects, such as electoral deepfakes.
AI is not destiny. If it increases the concentration, it will act as a corrosive resource. If it increases the productivity and prosperity of many, it will be the basis of a richer and more stable democracy.
Andrew Zanelato holds a bachelor’s degree in International Relations with an MBA in AI, Data Science and Big Data from PUCRS, Researcher in AI Governance and Livres associate.
Livres is a non-profit civil association that promotes liberal solutions for Brazil with a network of leaders, supporters and partners.
*This text does not necessarily reflect the opinion of Jovem Pan.
