Experts Warn ‘AI Bot Swarms’ on Social Media Pose Threat to Democracy

In a stark warning published today in Science, a prominent group of global experts has raised the alarm about a new threat to democracy: swarms of AI agents capable of mimicking human behavior to manipulate public opinion on an unprecedented scale.

The coalition, which includes Nobel Peace Prize winner Maria Ressa and leading researchers from prestigious institutions like Berkeley, Harvard, Oxford, Cambridge, and Yale, describes these “AI swarms” as a disruptive force that could soon infest social media and messaging platforms worldwide.

According to the experts, would-be autocrats could deploy these technologies to persuade populations to accept canceled elections or overturn legitimate results. The group predicts that such technology could be deployed at scale by the 2028 U.S. presidential election, though early versions have already been observed in 2024 elections in Taiwan, India, and Indonesia.

“These systems are capable of coordinating autonomously, infiltrating communities and fabricating consensus efficiently. By adaptively mimicking human social dynamics, they threaten democracy,” the authors warned in their joint statement.

The threat is particularly concerning because these AI swarms can precisely infiltrate online communities, learn their specific characteristics over time, and deliver increasingly convincing and carefully tailored falsehoods designed to shift public opinion on a mass scale.

Daniel Thilo Schroeder, a research scientist at Norway’s SINTEF research institute and one of the paper’s authors, emphasized how easily these systems could be deployed. “It’s just frightening how easy these things are to vibe code and just have small bot armies that can actually navigate online social media platforms and email and use these tools,” said Schroeder, who has been simulating such swarms in laboratory conditions.

The sophistication of these AI systems is rapidly advancing. They can increasingly adapt to human communication patterns, using appropriate slang and posting irregularly to avoid detection. Additionally, progress in “agentic” AI development means these systems can now autonomously plan and coordinate actions across multiple platforms.

Jonas Kunst, professor of communication at the BI Norwegian Business School and another author of the warning, explained the multiplier effect of these coordinated systems: “If these bots start to evolve into a collective and exchange information to solve a problem – in this case a malicious goal, namely analyzing a community and finding a weak spot – then coordination will increase their accuracy and efficiency. That is a really serious threat that we predict is going to materialize.”

Real-world examples are already emerging. In Taiwan, a frequent target of Chinese propaganda, AI bots have been increasingly engaging with citizens on platforms like Threads and Facebook over the past few months, according to Puma Shen, a Taiwanese Democratic Progressive Party MP who campaigns against Chinese disinformation.

These bots typically flood discussions with unverifiable information, creating what Shen describes as “information overload.” Some AI agents have been spreading narratives suggesting America will abandon Taiwan or encouraging younger Taiwanese to remain neutral in the China-Taiwan dispute by emphasizing its complexity.

“It’s not telling you that China’s great, but it’s [encouraging them] to be neutral,” Shen told the Guardian. “This is very dangerous, because then you think people like me are radical.”

The experts behind the warning include several prominent voices in technology and misinformation research, such as NYU’s Gary Marcus, a self-described “generative AI realist,” and Audrey Tang, Taiwan’s first digital minister, who has warned about authoritarian forces undermining electoral processes using AI.

However, not all experts are convinced that deployment will be as rapid or widespread as the warning suggests. Inga Trauthig, an adviser to the International Panel on the Information Environment, noted that politicians’ reluctance to cede campaign control to AI systems might slow adoption, adding that “most political propagandists I interview are still using older technologies and are not at this cutting edge.”

Nevertheless, Michael Wooldridge, professor of AI foundations at Oxford University, validated the overall concern: “I think it is entirely plausible that bad actors will try to mobilize virtual armies of LLM-powered agents to disrupt elections and manipulate public opinion… It’s technologically perfectly feasible.”

The authors are calling for coordinated global action to counter these risks, including the development of “swarm scanners” and watermarked content to help identify and combat AI-run misinformation campaigns before they can fundamentally undermine democratic processes.