“DeFi functions at high computational speeds, which means AI doesn’t really change the game,” remarked Kulechov. “Instead, it heightens a landscape that has always needed vigilant oversight.”
Nevertheless, Aave is observing that AI is revealing new types of vulnerabilities, including those that human auditors might have overlooked in the past.
“The Mythos research indicates that AI can identify longstanding bugs that were not previously prioritized,” he noted.
This wide-ranging impact remains crucial in a system where even minor vulnerabilities have the potential to erode trust or be exploited in larger attacks.
If attackers are able to act more swiftly, the pressing question is whether the security measures can adapt accordingly.
For both Gauntlet and Aave, the solution involves a fundamental shift in the security paradigm. Traditional audits conducted prior to deployment and ongoing monitoring were tailored for threats at human speeds. AI shortens that response time.
“To counteract aggressive AI, we must adopt an AI-centric strategy that prioritizes speed and ongoing adaptability,” stated Vijender from Gauntlet. This includes continuous auditing, real-time simulations, and systems designed with the expectation that breaches will occur.
A ‘greater way’
Aave has already incorporated AI into its processes, utilizing it for simulations and code assessments in conjunction with human auditors. “We embrace an AI-first strategy where it clearly enhances our efforts,” Kulechov mentioned from Aave Labs. “However, it serves to support rather than replace human-led audits.”