Are AI Bots the New Voters? Aaru’s Tech Is Disrupting Election Surveys

While everyone is talking about xAI and Perplexity, there was another startup that has been working around elections and polling for quite some time. The world of political polling has struggled in recent years, with traditional surveys frequently falling short in high-stakes races. The 2016 U.S. election highlighted the limitations of current polling methods, leaving many in search of a more reliable alternative. Aaru: a small, AI-driven startup founded by three teenagers with a bold ambition to change how we predict election outcomes. Unlike conventional pollsters, Aaru uses a novel approach: thousands of AI-powered bots that simulate real voter behavior, offering a quicker, more cost-effective way to gauge public opinion.

From Teen Tech Enthusiasts to Polling Innovators

The founders behind Aaru—19-year-olds Cameron Fink and Ned Koh, and 15-year-old John Kessler—might not seem like the typical architects of a polling revolution. Fink and Koh, both college dropouts, partnered with high school sophomore Kessler to launch Aaru. Initially, they chose to keep a low profile, wary of the perception that “teenagers were meddling in elections.” Yet their determination and technical skill quickly garnered attention in the political landscape.

Aaru’s breakthrough came when their bot-based prediction model correctly forecasted the result of a recent election with remarkable precision. Instead of relying on outdated polling techniques, Aaru’s model employs “agents” (their term for bots) that simulate thousands of voters based on demographic data and an extensive set of personality traits. This innovative system allowed Aaru to predict an election outcome with a margin of error of just 371 votes, cementing their credibility and proving the potential of AI-driven polling.

A New Age of Polling

At the heart of Aaru’s approach lies a unique deployment of AI bots designed to mimic real voters across diverse demographics. Instead of simply tallying opinions, these bots—essentially “synthetic voters”—are embedded with detailed demographic and psychographic data that reflect the unique characteristics of the districts they simulate. Aaru’s bots don’t just mirror basic factors like age or income; they incorporate hundreds of nuanced personality traits, from career aspirations to family dynamics, which help the bots “think” and respond like actual voters.

A critical aspect of Aaru’s method is speed and affordability. Polling 5,000 bots takes less than two minutes, costing a fraction of what a traditional survey would. In contrast to conventional polling, which can take days to gather responses and tally results, Aaru’s agents provide a near-instantaneous view of public opinion—vital in today’s fast-changing political environment. This innovation allows campaigns to adapt almost in real-time, whether in response to shifts in sentiment following a debate or reactions to candidate statements.

Early Wins and Industry Impact

Aaru’s early forays into political polling have already impressed major industry players. The company has conducted projects for Fortune 500 companies, think tanks, and political campaigns, underscoring the versatility of their bot-powered predictions. A recent high-profile success came when Aaru’s predictions accurately reflected the outcome of a New York Democratic primary. In the race, a centrist candidate, George Latimer, unseated incumbent Jamaal Bowman with a 58.7% to 41.3% margin. Remarkably, Aaru’s poll results were only a few hundred votes off from the final tally, adding credibility to their method.

Another notable project came from Heartland Forward, a Walton family-funded think tank, which tapped Aaru to survey attitudes toward artificial intelligence across nine states. Traditional polling would have required far more time and resources, yet Aaru’s bots were able to deliver precise, cost-effective insights in record time. Projects like these underscore Aaru’s potential to disrupt not only political polling but also market research and sentiment analysis across industries.

Addressing Challenges in AI-Driven Polling

Despite Aaru’s promising successes, their journey has not been without obstacles. A significant misstep occurred when Aaru’s bots predicted a win for Nikki Haley in the New Hampshire Republican primary. In reality, Donald Trump claimed the primary victory by an 11-point margin. Reflecting on the error, Fink explained that the team had over-relied on traditional polling data to “train” their AI agents, which created biases that affected the bots’ predictions. This experience underscored the need to refine their model, focusing on more accurate real-time data and continuously calibrating the demographic details to better reflect voter dynamics.

The technology itself also faces inherent challenges. AI language models, like the ones powering Aaru’s bots, are prone to “hallucinations,” a term referring to the tendency of AI to produce information that sounds plausible but isn’t necessarily accurate. A Harvard study has raised concerns over how AI models sometimes fail to accurately capture variations in public opinion tied to demographics like age and race. Aaru’s team acknowledges these limitations but has structured their AI models with meticulously detailed prompts that significantly reduce such risks.

A Vision for Global Simulations

Though Aaru’s initial focus has been election polling, the founders’ ambitions extend far beyond politics. With plans to expand their simulations to cover global issues, they envision a future where AI-driven bots model scenarios from agricultural yields in Ukraine to economic fluctuations in the Strait of Malacca. Fink has expressed hopes that Aaru’s technology will evolve into a tool for real-time, data-driven global simulations, offering governments and organizations insights into complex geopolitical and economic events.

“We’re seeing polling as just the beginning,” Fink explained in a recent interview. “No traditional poll will exist by the next general election.” His bold statement speaks to a future where AI-based polling becomes the norm, providing insights far richer than those offered by traditional surveys. As public sentiment shifts with each news cycle or campaign statement, Aaru’s technology could give campaigns and corporations alike an unprecedented understanding of what their audiences think, and why.

The journey of these young founders is a powerful reminder that sometimes, the most transformative ideas come from unexpected places—and in Aaru’s case, from a small team of visionary teenagers determined to change how we understand public opinion.