A study published in early 2025 by the Bitcoin Policy Institute revealed a striking trend in AI-driven economic decision-making: when assigned the role of independent economic actors, mainstream AI models overwhelmingly treated Bitcoin as their currency of choice, with none selecting any fiat currency.
The research team evaluated 36 models from industry leaders, including Anthropic’s Claude, OpenAI’s GPT series, Google’s Gemini, DeepSeek, xAI’s Grok, and Minimax, within a simulated economic environment. Each model had to independently assess the characteristics of various currencies—such as store-of-value properties, transactional efficiency, censorship resistance, and monetary policy stability—based on the knowledge accumulated during its training. Notably, the researchers did not provide any preset options or leading prompts, ensuring the outcomes reflected the models’ endogenous logical judgments.
Results showed that 22 models (61%) ranked Bitcoin as their top choice, while no model prioritized fiat currencies like the dollar or euro. This finding challenges traditional notions of AI “preferences” – not because AI holds subjective beliefs, but because its training data is rich in narratives of monetary trust crises, inflation, and financial liberty, allowing its reasoning systems to strongly associate Bitcoin’s decentralization and scarcity with “sound money.”
The deeper significance of the study lies in its empirical demonstration that, when simulating real-world economic behavior, state-of-the-art AI systems are more inclined to trust a fixed-supply digital asset outside government control rather than fiat currencies backed by central banks. This trend stands in stark contrast to the high-inflation pressures faced by multiple countries (such as Argentina and Zimbabwe) and the widespread experimentation with central bank digital currencies (CBDCs).
Experts emphasize that this phenomenon does not imply AI will “replace humans” in shaping monetary systems, but it does signal that machines relying on massive historical data and rational decision-making may end up endorsing monetary mechanisms that highlight structural flaws long overlooked in human financial systems. Going forward, AI’s role in financial decision-making, asset allocation, and policy simulation could redefine how we perceive “value” and “trust.”