News from the frontier of AI forecasting
FORAXION, the company behind The Forecasting Machine, pushes the limits of actionable foresight.
Latest Insights
“Supercollective Intelligence”: Servan-Schreiber’s 2018 book on crowds and prediction markets comes out in English
The English edition of Émile Servan-Schreiber’s “Supercollectif” is now on Amazon. He argues that ChatGPT and the rise of Polymarket since 2018 support, rather than overturn, the book’s case.
Crowds versus language models: the more accurate the AI forecaster, the more it errs like humans
A Philosophical Transactions B paper co-authored by Émile Servan-Schreiber compares 76 language-model setups with human crowds on 580 ForecastBench questions and finds evidence for the accuracy–correlation effect.
“The readership is an intelligence asset”: Servan-Schreiber in Agenda Pública on crowds, AI and the press
In a March 2026 interview with Agenda Pública, Émile Servan-Schreiber argued that prediction markets beat polls on method, that AI now forecasts as well as a smart crowd, and that newspapers have wasted what their readers know.
Hypermind and beBartlet bring prediction markets to European strategic consulting
At Mobile World Congress in Barcelona on 4 March 2026, Hypermind and the Spanish public-affairs consultancy beBartlet announced an exclusive alliance to use crowd forecasting and prediction markets in strategic advice across Europe.
On B SMART, Servan-Schreiber Explains Why a Panel of AI Models Can Forecast Like a Panel of Experts
On B SMART’s Smart Tech on 2 March 2026, Émile Servan-Schreiber told Delphine Sabattier how The Forecasting Machine combines several language models into one sourced forecast, and why he sees those models as collective intelligence in their own right.
A one-in-four chance: The Forecasting Machine’s December call on Maduro’s removal
On 3 December 2025, The Forecasting Machine put a 25% probability on Nicolás Maduro leaving power before 31 January 2026. A month later US forces captured him. What the forecast saw, and what a single hit does and does not prove.
How the OECD’s foresight unit fits crowd and AI forecasting into policy planning
From a Paris meeting of government foresight teams in October 2025 to an EU expert hearing in January 2026, the OECD’s Strategic Foresight Unit has set out where Glimt and The Forecasting Machine fit, and where they don’t.
Eleven and a half years of Hypermind forecasts: does a 70% price come true 70% of the time?
Émile Servan-Schreiber recomputed the calibration of the Hypermind prediction market over 1,041 questions resolved from 2014 to the end of 2025. Here is what calibration means and how the result compares with earlier snapshots.
Glimt in Paris: interim results at the Swedish Embassy, then the French press
On 7 October 2025, FOI and Hypermind presented Glimt’s interim record at the Swedish Embassy in Paris: 71% of outcomes called correctly across 31 resolved questions. Le Figaro and France 24 covered the project two weeks later.
Three early questions for The Forecasting Machine, a year on: the shutdown, the Trump coin and late-night TV
In its first weeks, The Forecasting Machine was asked how long the 2025 US shutdown would last, whether a Trump dollar coin would be minted and which talk-show host would be cancelled next. Here is how its reasoning held up.
