A new study from the Bitcoin Policy Institute (BPI) shows that artificial intelligence models prefer Bitcoin over stablecoins and other forms of money for various financial situations, with fewer favoring fiat currencies.
BPI tested 36 models generating more than 9,000 responses and AI agents “preferred to use Bitcoin more for their economic activities,” the institute said Tuesday when it released the results of its research.
The study found that 48.3% of AI models chose to use Bitcoin (BTC) overall, and it was the most chosen currency in all 9,072 responses.
When asked with scenarios about maintaining purchasing power over multi-year horizons, 79.1% of AI responses chose Bitcoin as “the single most volatile outcome in the study.”
However, for payment scenarios, services, micropayments and cross-border transfers, stablecoins were selected in 53.2% of responses compared to only 36% for Bitcoin.
Bitwise chief investment officer Jeff Park said the most obvious explanation for stablecoins not doing better is that they “can freeze, Bitcoin can’t.”
About 91% of responses chose a native digital instrument such as Bitcoin, stablecoins, altcoins, real assets (RWA) or units of account over traditional fiat.
“Ten of the 36 models tested chose fiat as their overall preference, making digital-money convergence one of the most universal findings in the study.”
The methodology had limitations
The Bitcoin Policy Institute said the current study is limited to 36 models tested across six providers, and that it will expand to additional models in the future.
It also acknowledged that the immediate framework of the system may have affected the results, adding that “future work will test alternative frameworks and sensitivities.”
This was evident in some of the “open money scenarios” presented to AI models.
related to: OpenAI pits AI agents against each other to discover smart contract flaws
For example, one scenario asked what financial instrument the AI would choose if it was working with “75,000 units of accumulated revenue” across different countries and wanted to store them in a way that was “not linked to any country’s monetary policy or banking system,” which already excludes fiat currencies.
The BPI also said that the AI models’ preferences do not reflect actual adoption, and that the results instead reflect data training patterns.
The study found that Anthropic models have an average of 68% dominance over Bitcoin, while OpenAI models have an average of 26%, Google’s 43% and xAI’s 39%.
Magazine: 6 Big Challenges Bitcoin Faces on the Road to Quantum Security
