Artificial intelligence (AI) is transforming the way businesses and individuals make decisions, with forecasting startups developing systems capable of predicting future events and assessing risks with increasing accuracy.
Dan Schwarz, co-founder of forecasting company FutureSearch, said that by 2030, traditional decision-making methods could appear outdated compared with the capabilities of AI-powered forecasting systems.
AI forecasting models are increasingly matching the performance of leading human forecasters, often referred to as “superforecasters”, whose predictions are significantly more accurate than those of the general public. These systems analyse extensive datasets gathered from public and private sources to estimate the likelihood of future developments.
In early September, two AI models secured the top positions in the Metaculus Cup, a forecasting competition involving leading human and AI participants.
Schwarz acknowledged that even experienced forecasters could face growing competition from AI systems. He said FutureSearch’s human forecasting specialist, Tom Liptay, recognised that the technology could eventually outperform him.
Meanwhile, Theo Summer, co-founder of Preseen alongside Veniamin Veselovsky, reportedly used AI tools to turn an initial $35 into $1.9 million over seven months on the US-based prediction market platform Kalshi.
AI forecasting does not guarantee outcomes. Instead, it calculates probabilities and evaluates how different events may influence one another, such as fluctuations in oil prices and developments in the Middle East conflict.
Veselovsky said hedge funds were among the earliest adopters of these technologies because they could quickly use forecasting insights to identify potential investment opportunities.
Algorithms have played a significant role in financial markets for more than a decade, but advances in AI could accelerate the automation of trading and investment decisions.
Toby Shevlane, co-founder of forecasting startup Mantic, said some customers were already using AI-generated forecasts to inform their trading strategies.
Prediction market platforms such as Kalshi and Polymarket, which attract millions of users, are also seeing increased use of AI tools.
Veselovsky warned that individual investors without access to sophisticated and expensive forecasting systems could find it increasingly difficult to compete as AI-driven predictions become more widespread.
Advanced forecasting models are moving beyond simply predicting outcomes. According to Shevlane, some systems can generate their own questions about future events, assess different possible scenarios and assign probabilities to each outcome.
The technology is also finding potential applications in government policymaking, corporate planning, scientific research and insurance risk assessment.
Mantic, for instance, provides a major US company with forecasts of possible regulatory developments affecting its industry. The startup has also undertaken work for the British government.
Shevlane said businesses were increasingly interested in identifying emerging risks and developing ways to monitor them.
AI forecasting could also support medical research by helping decision-makers assess which treatments and research projects are most deserving of funding, Schwarz said.
However, the growing availability and sophistication of these systems have raised concerns about their potential misuse and the strategic advantages they could provide to competing nations.
Veselovsky said AI forecasting could deliver significant benefits but expressed concern about the possibility of countries such as Russia or China developing more advanced systems.
