Contrary to grim forecasts, artificial intelligence has not become a catalyst for mass hacks in the decentralized finance sector. As my data shows, fears of total attack automation have not materialized. Dragonfly venture capital fund managing partner Haseeb Qureshi confirms this trend: despite a record number of incidents, the median damage in 2026 fell below $500,000, whereas a year earlier this figure stood at $2 million.
Why didn't AI become a weapon of mass destruction?
My analysis shows that attackers using AI are mainly targeting small or abandoned projects. This makes sense: major DeFi protocols have significantly strengthened their security over the past two years by implementing multi-level monitoring systems and automated response mechanisms. AI tools, such as generating phishing emails or automatically searching for vulnerabilities in smart contracts, have indeed become more accessible, but their effectiveness against well-protected platforms remains low.
Notably, the median attack size has decreased fourfold. This suggests that hackers have shifted from large, high-risk targets to smaller, less protected projects. This approach allows them to avoid close scrutiny from regulators and major security firms, but the profitability of such operations is significantly lower.
Expert's conclusion
The DeFi market demonstrates remarkable resilience in the face of technological threats. However, one should not be complacent: as AI tools become cheaper and more sophisticated, we may see an increase in attacks on medium-sized projects. Investors and developers should not relax but continue investing in security—this is what currently separates leaders from laggards in the DeFi world.