Artificial Intelligence puts democracy in check

Innovation will be the challenge every day to face the challenges that arise with the use of artificial intelligence (AI) in electoral processes, since there is no type of legislation or regulation in this regard; There is still no 100% reliable software to detect its use, when the massification of deepfakes can put democracies at risk and, begin to be taken into account, as factors to annul elections in an “extraordinary” way, by judicial authorities electoral.

This is what Everardo Bárcenas, a doctor in computer engineering from the UNAM and coordinator of the Artificial Intelligence Laboratory of that highest university, agreed to point out, the electoral advisor of the INE and member of the Complaints and Complaints Commission, Arturo Castillo, and the former magistrate of the Electoral Tribunal and advisor on the subject at the Council of Europe José Luis Vargas.

The challenge, commented Castillo and Vargas, will be to maintain fairness in the contest in the electoral processes and guarantee the free vote of citizens who are the central victims of the manipulation of information to try to distort their perception and influence their voting decision. , at a time when, according to Dr. Bárcenas, there is still no completely reliable software to detect the use of AI in identity theft.

THE PROBLEM OF AI

Dr. Bárcenas explained that it was the use of analysis with artificial intelligence that, in 2014 in the United States, influenced the presidential results, also the results of Brexit in Great Britain and “artificial intelligence is currently used a lot,” not only for data analysis, but also to “generate speech, communications, banners, posters, advertising” in campaigns.

He explained that the bots also use artificial intelligence to “pose themselves as real people and give their opinion about a candidate, they are biased, the idea is to generate a certain perception in some way and they are automatically, they are not people.”

The coordinator of the UNAM AI laboratory mentioned that modern bots can now not only publish opinions on social networks, but also engage in dialogues with users.

This makes it more difficult to identify these programs because they do it very efficiently. This is a challenge, identifying bots is much more difficult,” Bárcenas considered.

He said that it will be “complicated for the referee to do a census, to monitor, with these new technologies it seems to me that they are going to face a great challenge because the information is going to be much greater, technologically it is going to be very difficult to identify false, defamatory information, I am Thinking about videos that seem real, it will be difficult to audit them and see if what is being generated is real or not, that did not happen in other elections, and now it is going to be presented.”

Lagging legislation

For the advisor of the INE, Arturo Castillo, deepfakes began to be new in the world “three or four years ago, the legislation is not prepared to address it, nor is the technology, the technology we have is still not reliable enough to do a technical opinion on whether or not something is a deepfake.”

José Luis Vargas, former magistrate of the TEPJF, explained that phenomena such as identity theft are beginning to occur with “voices that do not come from a human being, but from artificial intelligence” and transmit messages that deceive citizens in a “viralized” way, which “It is not only an issue of a rumor or lie being released, but also the capacity to generate damage through viralization… it has to do with the famous echo chambers,” in which the Internet user only sees and hears content that he or she is familiar with. feels identified and is fed back with that same type of information.

The member of the INE Complaints Commission, Arturo Castillo, gave an example of Spain, where “they literally had to resort to experts, for example in voice recognition,” which in Mexico would be a complication in pre-campaigns and campaigns, because “if you have to resorting to experts and all that can take a lot of time and in campaigns, you don't have time to do this.”

CONTEST

Questioned about the possibility of viralized deepfakes that could break the fairness in the contest and the power of the Electoral Court to qualify and annul elections, former magistrate Vargas Valdés considered that, although the only grounds for annulment are clearly established in the law, such as illicit money in campaigns or the acquisition of time on radio and television, “abstract nullity never completely disappeared. When there are conditions that together lead to determining that the principles were violated, it is called annulment based on principles, when they are sufficient, which lead the body to determine that due to their severity and importance they are decisive for a result between first and second place, power "To appeal to that mixed bag, this hypothesis is not excepted, but it would undoubtedly be an extraordinary case."

THE IMMEDIATE SOLUTION

Vargas Valdés explains that the magistrates have had to “generate adjustments to the rules of the game through the interpretation of the rules. It is not ideal, what is ideal is regulation that addresses the right to information, personal data, various aspects and that legislators do that work so that judges apply the rules, that is not there, it is in its infancy, it does not exist yet.”

He mentioned that there are some attempts and in the international order the UN and UNESCO have made some efforts to integrate “multiple international efforts” to respond to these phenomena.

He added that, in the United States, Silicon Valley has tried to legislate as little as possible in that country, although President Biden has already begun to make proposals, while in the European Union, the Council of Europe has made greater progress in generating a “Law of Digital Governance” that seeks to establish similar parameters among all member countries to “counteract disinformation phenomena that affect democracies.”

SUPPLIERS

In Asia, China, South Korea, India, Vargas explained, countries “providers of an important part of artificial intelligence, there is no regulation other than what these states dictate, but there is still no interconnection at a global level.”

In Mexico, without anything in this regard and with the pre-campaigns already started, the INE advisor and member of the complaints commission, Arturo Castillo, expressed his personal opinion that they will have to “innovate”, explore the “possibility of using the characteristic features of a person as personal data, the voice, for example, is a biometric data, what happens if it is a biometric data that is subject to the protection of personal data, and in the same way as an authority can almost immediately limit the dissemination of advertising in which children or minors appear without authorization for its dissemination, you can do it very quickly in the same way", demonstrating that that voice or image in an audio or video is not authorized by the person, to order to remove the images. networks these messages.

He considered that the great challenge of digital technologies, in terms of the commission of crimes, “is the dispersion of responsibilities, they are dispersed not only at the territorial level, the crime is not committed by a person who is located in your geographical area, but that, furthermore, illicit potentials are committed by technology developed by a multiplicity of actors located in different parts of the planet.”

He also mentioned the importance of having collaboration agreements with social media platforms.

He recalled that the INE has already worked with Meta in the case of Facebook, Instagram and Whatsapp, the latter private messaging application, where the company already has mechanisms to "reduce the impact of viral content information."

THE INAI PREPARES

The National Institute of Transparency, Access to Public Information and Protection of Personal Data is already preparing criteria for the elections on issues of misinformation and misuse of personal data.

Commissioner Julieta Del Río considered that “there is going to be an impressive war, most of this process and we are prepared for that. For what? The parties, the INE and the candidates are obligated subjects” and although artificial intelligence is not regulated, “we would have to look into the violation of personal data and see the mechanisms, which Commissioner Josefina Román is already working on.” , to participate fully in these elections.

He reminded citizens and candidates that they can report any violation of personal data to the Inai, such as the use of AI to clone the image or voice of any political actor.

Here you can access news in real time

Know the most viral in Facebook Trending

Read the columnists of Excelsior Opinion

clm