For a few months now, American economist Daron Acemoglu, professor at the prestigious Massachusetts Institute of Technology (MIT), has been focusing on how to change the enormous concentration of power, resources and algorithmic capacity of Artificial Intelligence (AI).
The problem, for him, is that the technology companies that lead it — almost all of them in Silicon Valley, in the USA — are not only not willing to give up on this, but they care less and less about the fact that this dynamic affects democracy itself.
"They turned against liberalism, because they want to control information — and doing so, whether by a company or a government, is not compatible with liberalism," he tells BBC News Brasil.
This week, Nvidia, the main manufacturer of chips used by North American AI laboratories, saw its shares dry up by US$192 billion on the country's stock exchange after the revelation that the company's CEO and founder, Jensen Huang, had an extramarital relationship with a possible Chinese spy.
At the same time, Acemoglu seems convinced that Artificial Intelligence lives in a bubble close to bursting. According to his calculations, it will be necessary to pay 10% of the US GDP by 2032 to pay for the investments that are being made in the sector today.
"It's brain-blowing," he barks.
The economist — who became famous for having released Why Nations Fail: The Origins of Power, Prosperity and Poverty (Elsevier), in 2012, alongside economist James Robinson, from the University of Chicago (USA), with whom he shared the Nobel Prize for Economics in 2024 (the other laureate was Simon Johnson, from MIT) — spoke to BBC News Brasil days before the second round of the Brazilian presidential election.
Watch an excerpt of the interview in the video.
Also read the interview.
