Israeli startup 42AI has emerged from stealth with $62.5 million in seed funding to develop a behavioral world model designed to predict human decision-making. Founded after three years of development, the company is initially deploying its predictive system at financial brokerages to help forecast how individual traders respond to volatile market conditions. The startup combines psychological research, neuroscience, and genetics, utilizing datasets covering millions of individuals to map internal psychological tendencies into mathematical representations.
The transformer-based model analyzes various customer interactions, including support calls, onboarding processes, and transactions, to calculate the probability of a person making specific decisions under new circumstances. According to the company, the system aims to measure underlying motivations instead of simply extrapolating past actions, achieving an F1 score between 0.8 and 0.85 for predicting trading direction and volume. While the startup states that the model does not require personally identifiable information because it never knows the real identity of the individual, it constructs individual behavioral profiles using linked customer activity.
This behavioral world model matters because it introduces a novel approach to forecasting human probabilistic behavior in complex financial environments by converting psychological traits into machine learning data.