Unexpected Twists: How AI Sees the World of Finance

In this episode of "AI & Beyond," we delve into a groundbreaking global study comparing the financial decision-making of leading large language models—including GPT, Gemini, and DeepSeek—with real-world human choices from 53 countries. The research reveals that AI models tend to be risk-neutral, consistently prioritizing expected value in lottery scenarios—but occasionally faltering when weighing present versus future gains. Most unexpectedly, the collective financial logic of these AIs most closely mirrors the thinking patterns of participants from Tanzania, overturning assumptions that AI aligns with Western perspectives. This episode explores what drives these surprising cultural echoes in AI reasoning and spotlights the urgent need to understand and address hidden biases in financial AI systems. Tune in for an eye-opening look at the intersection of culture, technology, and intelligent decision-making.