Artificial intelligence is now demonstrating a remarkable and slightly unnerving ability to predict the nuances of human behavior and social dynamics. This was showcased in a global forecasting competition where a British AI successfully made predictions on events like a potential public feud between Donald Trump and Elon Musk, a task requiring more than just data analysis.
ManticAI’s eighth-place finish in the Metaculus Cup was built on its ability to forecast a range of 60 real-world events. Its success on questions involving human interactions is a sign that AI is moving beyond the predictable worlds of science and finance and into the less predictable realm of sociology and psychology.
To forecast a spat between two famously unpredictable billionaires, the AI cannot simply look at economic data. Its system must analyze their past behaviors, track their public statements and sentiment, and model how their personalities might clash under certain conditions. It’s a form of computational behavioral analysis.
Similarly, predicting the political future of a figure like Kemi Badenoch requires an understanding of party dynamics, public opinion, and political strategy. The AI is learning the “rules” of these human games by analyzing vast amounts of text describing how they have played out in the past.
While the AI lacks genuine consciousness or empathy, it is becoming an incredibly skilled observer and pattern-matcher of human social systems. Its growing prowess in this area opens up vast possibilities—from more effective marketing to improved diplomatic strategies—but also raises important ethical questions about the future of predictive social technologies.