Mantic, a London-based startup that claims to have built an AI system to predict world events, has emerged from stealth with a $4m (£2.97m) investment.
Having assembled a team of engineers from firms including Google DeepMind and Oxford and Cambridge, Mantic is working towards an AI system capable of what it calls “judgemental forecasting”.
The goal is to provide businesses and governments with forecasts of disruptions or significant landscape shifts.
Mantic said it will be able to predict events ranging from supply chain disruptions to geopolitical shocks.
The company offered an example query: Will Iran close the Strait of Hormuz before 2027?
A human ‘superforecaster’ – someone who calculates the probability of events based on statistical analysis – could attempt to answer this question based on the input of hundreds of factors.
For Mantic, this undeniably impressive skill has yet to truly inform the strategy of major corporations and governments, with concerns over reliability and accuracy.
So, the startup is automating the process in the hopes of producing forecasts that can guide major decisions.
The Camden-based startup was founded by Ben Day, who holds a PhD in machine learning from Cambridge and Toby Shevlane, formerly of Google DeepMind.
Mantic’s pre-seed funding round was led by Episode 1 Ventures with participation from DRW and angels.