La Jornada: UN panel of experts points out risks of artificial intelligence

At the first meeting of the United Nations Panel of Experts on Artificial Intelligence (AI), Yoshua Bengio, co-president of the organization, warned that the world does not have 15 years, as many think, to address problems such as the use of lethal autonomous weapons or the challenges of cybersecurity and the protection of key infrastructure for countries.

We have a vital task, he said, especially given the speed at which changes are taking place.

“We must provide a rigorous scientific basis and clear evidence, now,” he said.

Bengio, professor at the University of Montreal, Canada, and one of the most cited computer scientists in the world, in his message during the panel’s first in-person meeting, held in Madrid, Spain.

María Ressa, journalist and winner of the Nobel Peace Prize in 2021, also co-chair of the panel, recalled that in an “era of fracture, 193 countries agreed that the world deserves an independent and authoritative scientific panel that offers an explanation of what AI systems are, what they do and where they take us.”

And to carry out this task, he said, we have an independent scientific authority, both from governments, companies and institutions, which he described as radical in the current global context.

The meeting was also attended by 37 of the 40 scientific members of the panel, representing the five regions of the planet.

Ressa, a well-known critic of misinformation on social networks, described them as “the richest and most powerful failed state in the history of humanity,” which has generative AI as its engine, but will soon have agential AI, that is, interconnected generative artificial intelligence systems that operate with less human supervision to make decisions and perform specific tasks.

He warned that “every week, every day it becomes more powerful,” and stated that “democracy is being impacted, and it is done in three stages; the first is a narrative war, and is the engine of the entire model, lies repeated a million times become facts, and AI not only spreads these lies, it manufactures them on a large scale, adapts them to each of us and produces the synthetic consensus that we confuse with reality.”

The second stage, he said, is that “if you can’t distinguish fact from fiction, institutions fail, and that includes the press, the courts, checks and balances, and the third stage,” he added, “is what we are seeing in many places in the world, strategic corruption and kleptocracy.”

When citizens can no longer see clearly, and no one can hold power accountable, he indicated, everything is lost. Therefore, he highlighted the strategic role of the independent panel to provide scientific bases on AI and its impact on humanity.