Artificial intelligence swarms threaten democracy

The era of clumsy social media bots, which flooded the internet with obvious copy-and-paste patterns, is coming to an end. An international research team warns in the journal Science of a new level of escalation in the form of artificial intelligence swarms.

These are fleets of AI-controlled individuals who maintain a consistent identity and have digital memory. These agents do not act in isolation, but coordinate their behavior autonomously. In this way, they could create an artificial reality that is difficult to distinguish from human interaction.

See more articles from iGuRu.gr when you search for news on Google.

iGuRu.gr preferred source

Social data researcher David Garcia from the University of Konstanz, and co-author of the paper, describes these systems as highly adaptable. By merging large language models (LLMs) such as GPT, Gemini or Claude with multi-agent systems, “malicious AI swarms” are created that authentically mimic social dynamics.

They infiltrate groups, chat with real users, and react to events in real time. This chorus of seemingly independent voices creates the illusion of broad public consensus. In reality, it deliberately spreads misinformation.

The illusion of the majority as a psychological weapon

According to scientists, the danger lies less in isolated false reports and more in the gradual shift of social norms through an “artificial consensus.”

When users encounter a multitude of seemingly independent profiles expressing the same opinion, social pressure is created. This false impression that “everyone is saying it” massively influences people’s beliefs. Jonas Kunst from BI Norwegian Business School, also a co-author, sounds the alarm:

The basis of democratic discourse – independent voices – could collapse if a single actor controls thousands of Artificial Intelligence profiles.

According to the analysis, the threat is widespread:

In the long term, such swarms could manipulate the language, symbols, and identities of communities. In addition, there is the threat of “contagion” of the digital environment. As AI swarms flood the internet with false claims, this manipulated data flows into training future AI models. In this way, they indirectly extend their influence to established AI platforms. Studies suggest that such tactics are already being used in early forms.

The technological barrier is frighteningly low, as powerful language models are often accessible to everyone. Techniques such as “thought chain prompting” could be misused to construct chains of reasoning that closely resemble human ones. Other researchers have already shown that disinformation generated by AI is often rated as more trustworthy than human texts. Since these swarms require minimal supervision and operate across platforms, classic moderation of individual posts is doomed to fail.

So Garcia calls for change: it is important to investigate the collective behavior of large AI groups using methods from behavioral science. Only then will the risks arising from the interaction of many AI actors become apparent.

The sheer volume and variation of content simply overwhelms traditional fact-checking systems.

Strategies for a resilient digital democracy

Algorithms should be trained to detect statistically improbable coordination patterns. Another pillar would be distributed observation centers that collect AI influence data.

Additionally, verification options should be created for real users that protect data privacy but facilitate the distinction between humans and machines.

Ultimately, according to the group, financial incentives are also needed: revenue generation from fake interactions should be prevented and accountability for AI infrastructure managers should be increased.

Only through a combination of technical detection, independent monitoring, and regulatory guardrails can artificial swarms prevent the suppression of genuine diversity of views.

DOI: 10.1126/science.adz1697

newsbot

Although the press releases will range from very select to rare, I said I'd pass...because sometimes the editors hide.