OpenAI stated that it took down accounts belonging to an Iranian group, Storm-2035, for using its ChatGPT chatbot to generate content aimed at influencing the U.S. presidential election and other significant issues. This operation involved creating commentary on U.S. election candidates and sharing it on social media. Despite these efforts, the engagement on the posts was minimal, showing no meaningful audience interaction.
According to OpenAI, the majority of social media content produced by Storm-2035 received little to no likes or shares. Their investigation revealed that ChatGPT was manipulated to craft both long-form articles and concise social media comments, but little traction was gained, suggesting that the operation was ineffective in reaching or influencing audiences.
Earlier, a Microsoft threat-intelligence report highlighted that Storm-2035 consisted of four websites that posed as news outlets. This Iranian network was targeting U.S. voter groups across the political spectrum with polarizing messages about issues like the Israel-Hamas conflict and candidates in the presidential election. The aim was to sow division among voters.
OpenAI's actions to ban Storm-2035's accounts come amid heightened scrutiny over how AI can be exploited for influence and misinformation. The company's commitment to monitoring activities for potential policy violations reflects an ongoing concern about the undetected use of technology to manipulate public opinion during critical electoral periods.
Collection
[
|
...
]