In a remarkable innovation, 19-year-old entrepreneurs and college dropouts Cam Fink and Ned Koh have pioneered using AI chatbots to predict election outcomes. They notably forecasted the results of a high-profile New York Democratic primary without traditional human surveys. Their startup, Aaru, asserts that it has revolutionized election polling by employing thousands of AI agents, yielding more accurate predictions.
Their technology accurately forecasted centrist George Latimer’s victory over incumbent Jamaal Bowman with a precise margin of 58.7% to 41.3%. Aaru’s method utilizes census data to simulate voter behaviour, attracting interest from Fortune 500 companies, political campaigns, and think tanks. One California campaign now primarily uses Aaru’s technology for its polling needs.
Cam and Ned have designed their AI agents to emulate voter personalities, allowing them to dynamically absorb media and adjust their political preferences. Aaru’s system can generate a poll with 5,000 AI respondents in just one minute, significantly reducing costs compared to traditional surveys.
These AI agents are meticulously crafted with hundreds of personality traits and simulate real voters by constantly engaging with the internet, mirroring the media consumption patterns of the voters they represent. This process can lead to shifts in political preferences, as seen when many of Aaru’s AI agents initially shifted their support to Donald Trump following an assassination attempt.
Despite some skepticism regarding the reliability of large language models, Aaru’s successful predictions indicate a potential shift toward AI-based polling methods. The think tank Heartland Forward has already employed Aaru to survey AI perceptions across nine US states, signalling growing corporate interest in this innovative approach.
Cam Fink confidently predicts that AI chatbots’ predictive capabilities will soon render traditional polling obsolete, suggesting, “No traditional poll will exist by the time the next general election occurs.”
