According to ChainCatcher, Anthropic co-founder and CEO Dario Amodei told US lawmakers that open-source AI development is heading down a 'very dangerous path'. He argued that once a capable AI model is released as open source, developers lose the ability to effectively monitor how it is used, including the inability to track abuse, revoke access permissions, or dynamically update security safeguards. This creates increased risks. Compared with closed-source models, fully open models are much harder to govern securely and may lead to irreversible misuse.
The Dilemma of Open-Source AI Security Governance
During the hearing, Amodei emphasized that once an open-source AI model is released, developers cannot track where the model is used, who modifies it, or for what purpose. This lack of transparency directly undermines security defenses: when the model is maliciously used to generate harmful content, automate cyberattacks, or design biological weapons, developers have almost no ability to cut off the source of risk in time. In contrast, closed models (such as ChatGPT or Claude) can limit abuse through API permissions, terms of use, and real-time monitoring, while open-source models rely entirely on the goodwill of downstream users—something nearly impossible to enforce globally.
Potential Implications for the Crypto Industry
Although Amodei's warning primarily targets general AI safety, its logic extends to the crypto space. The cryptocurrency industry increasingly relies on AI for smart contract auditing, on-chain transaction analysis, and automated market making. If an ungoverned open-source AI model is maliciously programmed or data-poisoned, it could directly lead to DeFi protocol attacks, oracle manipulation, or even the creation of entirely autonomous malicious smart contracts. Furthermore, decentralized AI projects (e.g., blockchain-based AI training networks) that adopt open-source models face similarly irreversible security risks—once model parameters are tampered with, on-chain governance mechanisms often cannot respond and patch as quickly as centralized APIs.
Amodei's congressional testimony serves as a wake-up call on the trade-off between openness and security. For the crypto industry, finding ways to preserve the ethos of decentralization while introducing effective AI safety governance frameworks is becoming an increasingly urgent challenge.