The Impact of Anthropic's Mythos Model on Crypto Industry Security

The introduction of Anthropic's Mythos AI model has sparked a significant shift in the crypto industry's approach to security. For years, decentralized finance has focused on defending smart contracts through code audits and vulnerability assessments. However, Mythos, designed to identify and exploit weaknesses across systems, is pushing the industry to look beyond code and into the underlying infrastructure. According to Paul Vijender, head of security at Gauntlet, a risk management firm, the greater risks lie in the infrastructure, including key management systems, signing services, bridges, oracle networks, and cryptographic layers. These components are often less visible and outside traditional audit scope. The recent security breach at web infrastructure provider Vercel, which may have exposed customer API keys, highlights the importance of reviewing code and rotating credentials. Vercel attributed the intrusion to a compromised Google Workspace connection via a third-party AI tool. Mythos belongs to a new class of AI systems that simulate adversaries, exploring how protocols interact and testing small weaknesses that can be combined into real-world exploits. This approach has drawn attention beyond crypto, with banks like JP Morgan treating AI-driven cyber risk as systemic and exploring tools like Mythos for stress testing. Early findings from models like Mythos have identified weaknesses in the behind-the-scenes systems that keep crypto platforms secure, including key protection technology and inter-system communication. Vijender believes AI models are especially valuable in identifying multi-step exploit chains and infrastructure-layer vulnerabilities that traditional audits often miss. The shift in focus matters in a system built on composability, where DeFi protocols interconnect and share liquidity, relying on common oracles and layers of integrations. This interconnectedness has driven growth but also creates pathways for risk to spread, as seen in recent bridge exploits. Without AI, tracing dependencies is challenging, but with AI, they can be mapped and exploited at scale, resulting in a shift from isolated exploits to systemic failures that cascade across protocols. Some industry leaders, like Stani Kulechov of Aave Labs, see Mythos as an acceleration rather than a turning point, reflecting the dynamics already at play in DeFi's adversarial environment. AI models represent an evolution in the tools used to achieve exploits, and DeFi is already built for machine-speed attacks. However, Aave is seeing AI surface new categories of vulnerabilities, including issues that human auditors may have previously deprioritized. The breadth of AI-driven threats still matters in a system where even smaller vulnerabilities can undermine trust or be combined into larger exploits. To defend against offensive AI, Gauntlet and Aave advocate for changing the security model itself, adopting an AI-centric approach with continuous auditing, real-time simulation, and systems built with the assumption that breaches will happen. Aave has integrated AI into its workflows, using it for simulations and code review alongside human auditors. Ultimately, the long-term effect of AI on the crypto industry may be less disruption than divergence, with projects that prioritize security having a greater ability to test and harden systems before launching, while those that don't will be most at risk.