One of the greatest challenges in forecasting is the forecaster themselves. Human beings are riddled with cognitive biases—from wishful thinking to political leanings—that can unconsciously skew our predictions. The recent success of a British AI in a global competition highlights the promise of a more objective, less biased approach to seeing the future.
ManticAI, the system that placed eighth in the Metaculus Cup, has no hopes, fears, or political affiliations. Its predictions are the product of data and algorithms, not emotion or ideology. This inherent objectivity is perhaps its single greatest advantage over human competitors.
This was explicitly noted by ManticAI’s co-founder, Toby Shevlane. He pointed out that human forecasters often “cluster around the community average,” a form of social bias. The AI, in contrast, “often strongly disagreed,” taking positions based solely on its own analysis. Its high ranking proves this objective approach is highly effective.
This promise of objectivity is a game-changer. Imagine a world where economic forecasts are free from political spin, or where intelligence assessments are not influenced by pre-existing policy preferences. AI could serve as a neutral arbiter, providing a baseline prediction grounded purely in evidence.
Of course, AI is not entirely free from bias, as it can inherit biases from the data it’s trained on. However, the system’s design—using multiple models to challenge each other—and its freedom from human psychological flaws represent a significant step toward a more objective form of analysis. In a world awash in misinformation and biased narratives, the clear-eyed perspective of an AI could be an invaluable resource.