AI Forecasting Startups Say Their Tools Now Match the Best Human Predictors

By

October 12, 2026

A group of startups say artificial intelligence can now predict future events as well as the best human experts, IOL reported. Their tools are used to forecast things like stock prices, oil prices and even how songs perform on the charts.

In early September, two AI models took the top places in the Metaculus Cup. The cup is a contest for prediction champions, and both humans and machines can enter.

“By 2030, we’re going to feel that what we were doing for our decision-making was complete Stone Age thinking,” said Dan Schwarz, cofounder of the forecasting company FutureSearch.

According to the report, AI models now perform about as well as so-called superforecasters. These are people whose guesses about the future are much better than those of most people. The models are trained on large amounts of public and private data.

FutureSearch employs its own superforecaster, Tom Liptay. Schwarz said Liptay knows he may not be safe in his job in the medium term. “He knows that these things are going to surpass him,” Schwarz said.

Theo Summer cofounded the startup Preseen with Veniamin Veselovsky. The report said Summer used AI tools to turn $35 into $1.9 million in seven months. He did this on Kalshi, an American prediction website.

The tools do not promise a result. Instead, they give the chance that something will happen. They can also check how linked events affect each other, such as oil prices and the war in the Middle East.

Veselovsky said hedge funds are usually the fastest buyers because they move quickly when they see a chance to make money. He said ordinary investors without costly AI tools may soon lose out on sites like Kalshi and Polymarket. He said these sites “will soon just be dominated by AI forecasting systems.”

Toby Shevlane, cofounder of another startup called Mantic, said the tools already affect markets in a small way. “We already have customers who use it when making trading decisions,” he said.

Shevlane said the newest models can also write their own questions about the future. They then give a chance to each possible outcome.

The report said the tools are starting to be used outside finance. Possible uses include government policy, company planning, science research and insurance risk.

Mantic gives a large American company likely scenarios for new rules in its industry. It has also done work for the British government. Companies “are interested in risks and how to monitor them,” Shevlane said.

Schwarz said other customers may want to use the tools in medical research. This could help them decide which treatments are best to fund.

Veselovsky said he also has worries about the technology. “I think that good can be derived from this,” he said, but “I’m also a little bit scared because, you know, maybe Russia or China has better systems.”

Source: this article is based on reporting by IOL Technology. Image: IOL Technology. Written with the help of AI and published by the Tzaneen Voice Technology desk. See our Editorial Standards.