” Artificial intelligence (AI) is risk to democratic elections“warned the Deputy General Director Social Sciences and Humanities UNESCO, Mexican Gabriela Ramosin an interview with EFE in Nairobi.

“We know this because of the accelerated spreaddeep fakes(ultrafalse) due to increasing systems Generative AI that enhance their analytical capabilities and results,” said Ramos, who attended the East African Sub-Regional Forum on Artificial Intelligence in the Kenyan capital this week.

Spread of fake news

Ramos admitted that there is “spreading false information“, which affects elections in the countries around worldas well as the US presidential elections, which will be held on November 5th.

However, this North American country “is I’m thinking about banning “deep fakes”which, in all likelihood, are genuine fake reproductions in audiovisual or written format,” the deputy director general said.

The UN agency has employment program So the platforms that work with AI have a responsibility.

Also with Ethical guidelines According to the UNESCO document on artificial intelligence, signed in 2021 by 194 countries, “concrete actions are being taken,” he noted.

“If you was finesif you turn it into crime“You’re increasing the cost to those who are going to use these types of techniques to influence or negatively impact an election,” he added.

Ramos emphasized that not only legislative decisions to ensure good use of this technology, but also technical solutions which imply, for example, “that any image, voice, video created by artificial intelligence of real people have watermark“.

These types of tools and technical standards will allow users to identify false information.

Legal Answers

He also called for a legal response, recalling that Mexico’s new president, Claudia Sheinbaum, was victim during the election campaign of the city. fake video in which she invited people to invest in the supposed financial platform.

“Necessary legal mechanism accountability and compensation regimes established within the rule of law, so that the elected president of Mexico in this case can appeal to a court that trained For chase this is a crime and demand responsibilities“he explained.

If billions of dollars are invested in software companies like Open AI to launch the Chat GPT AI platform, he insisted, governments should invest V institutional capacity, analytics And staff to protect society.

On the other hand, the head of UNESCO stressed that before bringing these products to market, companies must take responsibility from the very beginning.

“Is the data representative? Is it biased? Is it qualitative? What about your algorithms? What kind of conceptual framework are you using?” he asked about the frameworks to consider.

In his opinion, this new technological trend should be regulated from all sides, including the working conditions of content curators.

“In Kenya, Chat GPT has hired workers who earn $2 an hour to identify harmful content “in a statement,” Ramos said, stressing that these mechanisms to curb negative reporting “cannot rely solely on workers in developing countries who are subject to such cruelty.”

The Impact of AI in Latin America

However, AI is capable have a positive impact in societies, the deputy director emphasized, saying that health sector This is a potential area for the establishment of teleconsultations using this technology in regions such as Latin America.

“This could mean support tool “It’s very accurate for health workers to diagnose” and to access areas where there are no hospital resources, he suggested.

Also could be an ally solve other problems such as “problem corruptionwhich is so present throughout the continent,” the deputy director said, referring to the Latin American region.

Or cut discrimination against womenadded Ramos, who “proudly” recalled that “Mexico (…) has the first woman to rule in the history of the country.”