OpenAI announced on Friday that it had taken down the accounts of an Iranian group using its chatbot, ChatGPT, to create content with the aim of influencing the U.S. presidential election and other important issues.
Dubbed “Storm-2035,” the attack involved the use of ChatGPT to generate content related to various topics, including discussions on the U.S. presidential election, the Gaza conflict, and Israel’s involvement in the Olympics. This content was then shared on social media platforms and websites.
A Microsoft-backed AI company investigation revealed that ChatGPT was being utilized to produce lengthy articles and short comments for social media.
OpenAI noted that this strategy did not result in significant engagement from the audience, as most of the social media posts had minimal likes, shares, or comments. There was also no evidence of the web articles being shared on social media platforms.
These accounts have been banned from using OpenAI’s services, and the company stated that it will continue to monitor them for any policy violations.
In an early August report by Microsoft threat intelligence, it was revealed that an Iranian network called Storm 2035, operating through four websites posing as news outlets, was actively interacting with U.S. voters across the political spectrum.
The network’s activities focused on generating divisive messages on topics like U.S. presidential candidates, LGBTQ rights, and the Israel-Hamas conflict.
As the November 5th presidential election approaches, the battle between Democratic candidate Kamala Harris and Republican opponent Donald Trump intensifies.
OpenAI previously disrupted five covert influence operations in May that attempted to use their models for deceptive online activities.
Source: www.theguardian.com
