AI Models Prefer Bitcoin, Study Finds
A new study by the Bitcoin Policy Institute found that AI models prefer digital money over traditional fiat currency. Researchers tested 36 frontier AI models, concluding they would choose to transact with currencies like Bitcoin if operating as autonomous economic agents.
The study from the Bitcoin Policy Institute evaluated 36 different AI models from six prominent providers, including OpenAI, Google, and Anthropic. It presented these models with 9,072 neutral monetary scenarios to determine their preferences without leading prompts. The results showed a strong inclination for digital money, with over 90% of the AI responses favoring it over traditional fiat currencies. When the choices were broken down, Bitcoin was the top selection in 48.3% of all responses. The preference for Bitcoin was particularly strong in scenarios related to long-term value storage, where it was chosen 79.1% of the time. This suggests that AI models, when considering economic principles, identify Bitcoin's potential as a resilient asset for preserving value. For everyday transactions and payments, however, the AI models showed a preference for stablecoins, selecting them 53.2% of the time compared to Bitcoin's 36.0%. This indicates a nuanced understanding of different digital assets, favoring the price stability of stablecoins for frequent, smaller transactions while seeing Bitcoin as a form of "hard money" for savings. Interestingly, some AI models independently proposed their own forms of currency without any prompting. These suggestions were often based on units of energy or computing power, such as kilowatt-hours or GPU-hours, reflecting a transactional framework rooted in the resources that are fundamental to AI operations. The preference for digital currencies like Bitcoin among AI is linked to their inherent properties. Cryptocurrencies are permissionless, meaning an AI agent would not need to establish an identity to use them, unlike the traditional banking system. This is a critical feature for autonomous agents that would need to transact without direct human oversight. The concept of "programmable money" is also a key factor in the intersection of AI and blockchain. Digital currencies can be programmed with self-executing logic, allowing for automated and complex financial transactions without intermediaries. This aligns with the operational nature of AI agents, which could autonomously manage digital wallets and execute transactions based on predefined conditions. Looking ahead, the development of AI agents as economic actors could have significant implications for financial infrastructure. As AI-driven systems become more integrated into the economy, the demand for a native, digital, and automated form of money is expected to grow, creating a potential pathway for the wider adoption of cryptocurrencies.
