Bad actors are using OpenAI's platform to disrupt elections, but with minimal viral engagement.

• ChatGPT is being targeted by threat actors to manipulate election results, according to a report by OpenAI on Wednesday.
• OpenAI has disrupted over 20 operations and deceptive networks globally that attempted to use its models since January.
• This year's elections will impact over 4 billion individuals across 40 nations.

Across the world, cyber actors are increasingly choosing OpenAI as their preferred platform for influencing democratic elections.

The ChatGPT creator stated in a 54-page report published Wednesday that its models have been disrupted by more than 20 operations and deceptive networks worldwide, with threats ranging from AI-generated website articles to fake social media posts.

The company's update on "influence and cyber operations" aimed to provide a "snapshot" of its observations and identify an initial set of trends that can inform discussions on how AI fits into the broader threat landscape.

The report from OpenAI about the impact of AI on elections is released just before the US presidential election, which takes place in a significant year for global elections, affecting over 4 billion people in more than 40 countries. The increasing use of AI-generated content has raised concerns about election-related misinformation, with the number of deepfakes created increasing by 900% annually, according to data from Clarity, a machine learning firm.

The spread of false information in elections has been a persistent issue, with Russian actors exploiting social media during the 2016 U.S. presidential campaign and social networks being flooded with misinformation on Covid vaccines and election fraud in 2020.

The current preoccupation of lawmakers is centered on the growing prevalence of generative AI, which gained momentum in late 2022 with the release of ChatGPT and is now being utilized by businesses of all scales.

The report by OpenAI stated that the use of AI in election-related tasks varied in complexity, from basic content generation to intricate processes to analyze and respond to social media posts. The content primarily focused on elections in the U.S. and Rwanda, with a lesser emphasis on elections in India and the EU, according to OpenAI.

In August, an Iranian operation used OpenAI's products to generate articles and social media comments about the U.S. election and other topics, but the majority of identified posts received little engagement. In July, OpenAI banned ChatGPT accounts in Rwanda that were posting election-related comments on X. In May, an Israeli company used ChatGPT to generate social media comments about elections in India. OpenAI was able to address the cases within less than 24 hours.

In June, OpenAI acknowledged a secret operation that utilized its technology to produce remarks regarding the European Parliament elections in France and U.S., German, Italian, and Polish politics. The company stated that although the majority of social media posts it identified received minimal likes or shares, some actual individuals responded to the AI-generated posts.

The company stated that none of their election-related operations were able to achieve "viral engagement" or "sustained audiences" through the use of ChatGPT and OpenAI's other tools.

WATCH: Outlook of election could be positive or very negative for China