Ethereum co-founder Vitalik Buterin has issued a cautionary message to developers regarding the integration of artificial intelligence (AI) with blockchain technology. In a blog post on January 30, Buterin expressed his concerns, urging developers to exercise caution, especially when deploying new applications in high-value and high-risk environments. He highlighted potential risks, such as the vulnerability of AI oracles, emphasizing the substantial financial losses that could occur if a hack were to compromise these oracles.
Buterin acknowledged the potential of AI in micro-scale prediction markets, suggesting that AI could participate in these mechanisms in ways that are impractical for humans. He also proposed using AI in crypto wallet interfaces to enhance user understanding and interaction by interpreting transactions, signatures, and detecting potential scams. While expressing reservations about purely AI interfaces due to increased risk, Buterin emphasized that AI could effectively complement traditional interfaces.
In Buterin's perspective, the riskiest application is using AI to enforce rules or governance in cryptographic systems, as open-source AI models may be susceptible to adversarial attacks. He contrasted this with closed-source AI models, citing the security of obscurity they provide. However, he noted the lack of transparency and guarantees of impartiality in closed-source models.
Buterin cited Worldcoin, a crypto startup associated with OpenAI, as an example of closed-source AI that relies on trusted hardware rather than allowing open access to AI models. He highlighted the challenge of creating decentralized AI using blockchain and cryptocurrencies, particularly in developing a single decentralized trusted AI that various applications can leverage. Buterin acknowledged the potential benefits of such applications in avoiding centralization risks but cautioned that underlying assumptions could fail in various ways.
