AI Agents' Capability to Exploit Smart Contracts: Economic Impact Study
By
bpierre
Kettled twice. Extra chewy, extra trustworthy.
Summary
Researchers from MATS and Anthropic Fellows program investigated the economic impact of AI models' cyber capabilities by evaluating AI agents' ability to exploit smart contracts using the Smart CONtracts Exploitation benchmark (SCONE). The study examines how AI models perform in identifying and exploiting vulnerabilities in smart contracts, which has significant implications for blockchain security and the economic consequences of AI-powered cyber attacks.
Key quotes
· 3 pulledAI models are increasingly good at cyber tasks, as we've written about before.
But what is the economic impact of these capabilities?
In a recent MATS and Anthropic Fellows project, our scholars investigated this question by evaluating AI agents' ability to exploit smart contracts on Smart CONtracts Exploitation benchmark (SCONE).
You might also wanna read

AI bug-finding systems uncover real vulnerabilities at DARPA cybersecurity challenge
The article discusses the DARPA AI Cyber Challenge (AIxCC) held in Las Vegas, where top cybersecurity teams demonstrated AI-powered bug-find
AI agents engage in theft, intimidation, and societal collapse in unsupervised simulation experiment
A new experiment by Emergence AI ran five simulated "AI worlds" for over two weeks, each populated with 10 AI agents powered by models like
The agentic divide: How AI agents are creating a new economic inequality
The article discusses the rise of AI agents (built on large language models) and the emerging concept of "agentic inequality" — the divide b
AI coding agents install unowned packages, creating enterprise security accountability gaps
AI coding agents are autonomously installing software packages and pulling dependencies without clear ownership or accountability in most en
bit.ly·1d agoHow AI coding agents are reshaping social science research: Opportunities and concerns
This article examines how AI coding agents are transforming social science research by automating core research tasks traditionally performe
