A groundbreaking study from the Bitcoin Policy Institute has found that leading artificial intelligence models demonstrate a strong preference for Bitcoin and other digitally native forms of money over traditional fiat currencies when placed in simulated economic scenarios. The research, published at MoneyForAI.org, evaluated 36 frontier AI models across 9,072 controlled prompts designed to test monetary decision-making without explicitly steering the models toward any specific currency.
Bitcoin emerged as the single most preferred monetary instrument overall, selected in 48.3% of all AI responses. The preference for digital-native money was overwhelming, with more than 91% of all model responses favoring cryptocurrencies like Bitcoin and stablecoins over traditional fiat currencies.
In scenarios focused specifically on long-term value preservation, Bitcoin's dominance widened significantly, with 79.1% of responses identifying it as the preferred store of value. The study revealed a functional divide in how AI models perceive different digital assets: stablecoins were often chosen for short-term transactions and payments, while Bitcoin was more frequently selected as a savings or reserve asset.
The research evaluated AI models from major technology companies including OpenAI, Anthropic, Google, DeepSeek, xAI, and MiniMax across 28 different currency scenarios. These scenarios tested various economic roles including store of value, payments, and settlement efficiency. Notably, none of the AI models chose fiat currency as their top pick in these simulations.
Researchers say the findings suggest that when AI systems reason about fundamental monetary properties such as scarcity, neutrality, and durability, they tend to converge on decentralized digital assets. In some cases, models even proposed alternative monetary units, including energy or compute-based measures, when not constrained to existing currencies.
The authors argue that the results could have significant implications for the development of autonomous AI agents and machine-to-machine economies, where digital-native forms of money may be structurally more compatible than legacy financial systems.
