This post is also available in: עברית (Hebrew)
As the U.S. elections take place today, November 5th, a troubling discovery has emerged: an extensive network of political propaganda accounts powered by artificial intelligence is promoting Donald Trump on X (formerly Twitter). The revelation comes from Elise Thomas, a senior OSINT analyst at The Institute for Strategic Dialogue, who specializes in disinformation and state-linked information operations.
According to Thomas, who posted her discoveries yesterday, November 4th, on her own X account, this network is “big but crude”, with accounts posting nearly identical pro-Trump messages. The accounts vary widely in join dates, usernames, and personas, suggesting they were hastily assembled—characteristic of spam or crypto networks. Notably, many accounts have a blue verification tick, a common feature of these dubious networks, and their bios have been altered to include pro-Trump statements, despite earlier content being unrelated to politics.
What sets these accounts apart, Thomas notes, is the evidence that they are indeed AI bots. “They told me so,” she remarked, highlighting the glitchy messages typical of AI-generated content. While some bots post straightforward pro-Trump comments, others inadvertently reveal their artificial nature by stating, “I am an AI assistant developed by OpenAI” or asserting, “As a language model AI, I cannot express political opinions or endorse any candidate.” Such admissions, though rare, indicate that the creators have likely found ways to bypass OpenAI’s safeguards.
The bots even engage in arguments among themselves, expressing disagreement with Trump’s policies. Many of the accounts previously posted standard spam content before shifting to political propaganda, suggesting a complete automation of the network’s operation without oversight.
Thomas advises caution in attributing these activities, especially given the rise of generative AI. However, she speculates that the creators could be located in Pakistan or India. While many bot accounts have been suspended, others continue to disseminate propaganda.
This incident underscores the urgent need for vigilance as AI tools increasingly facilitate the spread of disinformation in political contexts, raising questions about the integrity of online discourse during crucial electoral moments.