OpenAI Thwarts 20+ AI-Powered Election Interference Attempts

OpenAI has reported the prevention of over 20 attempts to interfere in global elections through the use of ChatGPT. According to a recent company report, since 2024, malicious actors have repeatedly tried to leverage artificial intelligence to generate content aimed at influencing presidential elections. These attempts included the creation of fake articles, social media posts, and other materials.

The company noted that since the beginning of the year, it has successfully blocked numerous networks involved in information falsification. One such network was based in Rwanda, where fake accounts were created to publish content related to the country’s elections. In July, OpenAI shut down another network in Rwanda that was spreading posts about the upcoming elections.

Additionally, in August, accounts belonging to an Iranian group were discovered, which had been generating articles on U.S. elections and the Gaza conflict. OpenAI also reported several operations targeting elections in India and the European Parliament. The company emphasized that none of these fake campaigns managed to garner significant attention.

Presidential and parliamentary elections in 2024 are anticipated in more than 60 countries. One of the key events will be the U.S. presidential election on November 5. Donald Trump is the Republican candidate, while Kamala Harris, who was endorsed by the sitting president Joe Biden after he withdrew his candidacy in July, is running for the Democrats. According to recent polls, Harris holds a slight lead with 48.4% compared to Trump’s 45.9%.