Foreign governments are using ChatGPT to influence voters, report finds

As Election Day approaches, public agencies and private companies have increased efforts to stop the possible influence of foreign governments on the Latino vote.

According to a statement from OpenAI, the company deactivated a series of Iranian government accounts that allegedly used ChatGPT to promote false content about the U.S. presidential election on social media.

In addition, OpenAI reported that the Storm-2035 covert operation also generated false content in Spanish and English "about politics in Venezuela (and) the rights of Latino communities in the United States," among other topics.

"Our investigation revealed that this operation used ChatGPT for two purposes: to generate long articles and shorter comments on social media," says the statement published in August.

One comment posted on X mentioned a possible increase in the cost of immigration procedures if Kamala Harris is elected president.

News4 asked the Harris-Walz campaign about this matter, and we have not yet received a response.

However, attempts by foreign governments to influence US presidential elections are not new.

In July 2019, the Senate Intelligence Committee published the first part of the report of a bipartisan investigation that studied Russia's attempt to influence the 2016 election.

Also, the final report by special counsel Robert Mueller revealed that US intelligence agencies assessed that "social media manipulation was part of a Russian intelligence operation designed to undermine American democracy by exacerbating divisions, harming Democrat Hillary Clinton and helping Trump," NBC News reported in 2019.

At that time, the Senate Intelligence Committee recommended that the federal government provide the necessary resources and information to local governments to detect possible cases of electoral interference. The panel also indicated that "the United States must create an effective deterrent" to combat these efforts.

ARTIFICIAL INTELLIGENCE AND THE LATINO VOTE

Almost five years after the publication of these reports, OpenAI detected that the Iranian government published propaganda in Spanish, with content alluding to the Latino community.

But why would foreign governments want to influence the vote of Latinos specifically?

"Let's say that Latinos could be a door that they can open to generate distrust in the community and incite protests. That's something that happens. It's not really about AI, but it's more like a geopolitical movement based on what Russia has done," the journalist Cristina Tardaguila explained.

The researcher at the Institute for Digital Democracy of the Americas also clarified that the impact of artificial intelligence (AI) on the 2024 elections is not very significant.

In fact, OpenAI added in its report that the impact of the posts in question "does not seem to have achieved significant participation in the audience." However, Tardaguila acknowledged that using this technology has promoted fear in some groups.

"'AI is going to destroy our democracy.' 'It is destroying our elections.' That is something that has no roots in real data, but it is much easier to generate that emotional feeling of fear," she added.

The director of the US Cybersecurity Agency, Jen Easterly, confirmed what Tardaguila said: U.S. Elections are safe from these attacks.

Easterly told The Associated Press that foreign governments will not be able to alter the result of this year's elections thanks to the infrastructure in place to secure elections and ballot counting.

Several DMV residents told Telemundo 44, our sister station, how they verify the truth of social media posts.

"Sometimes I check," José Troche said in Spanish. "If the information is so exaggerated, unbelievable, logical, yes. I also want to see where it comes from to see where the rumor started and how it got to the level where it is now."

"You have to have the desire to search and that's hard for a lot of people because they don't do it," he concluded.

HOW TO SPOT A "TROLL" ON X

A report by the cybersecurity firm Symantec revealed that, when Russia interfered with the election in 2016, the Internet Research Agency (IRA) operation used "trolls" to promote propaganda on the platform formerly known as Twitter.

"The IRA's basic strategy was to use a small core of Twitter accounts to spread new content and leverage a broader set of automated accounts to amplify those messages," NBC News reported in 2019.

When an account is not real, it is called a "troll," meaning it is a profile usually created to promote misleading information.

If you want to learn how to spot fake accounts, there are several online tools that can help.

One of them is called Spot the Troll. The website, created by Clemson University in South Carolina, features eight accounts, some made by Russia. It uses them as an example to teach the audience how to spot "trolls" on the social network now known as X.

Click here to see what it's all about.