Ari Juels: Can AI Weaponize Blockchain Smart Contracts? - Chainlink
Nov 19, 2024
auto_awesome
Ari Juels, a Cornell professor and chief scientist at Chainlink, dives into the potential perils when AI meets blockchain. He warns about rogue smart contracts that could be weaponized for criminal purposes and discusses their convoluted relationship with large language models. Juels also tackles the evolving landscape of decentralized organizations, ethical considerations in governance, and the importance of transparency to prevent malicious activities. Plus, he elaborates on innovative metrics like 'voting block entropy' that could reshape DAO performance assessments.
The intersection of AI and blockchain raises concerns about rogue smart contracts being weaponized for criminal activities, necessitating caution and regulation.
Oracle systems are crucial for integrating AI capabilities into smart contracts, allowing them to access off-chain data and enhance their functionality.
Decentralized Autonomous Organizations (DAOs) require innovative metrics like voting block entropy to more accurately measure true decentralization beyond traditional token distribution.
Deep dives
Understanding Rogue Smart Contracts
Rogue smart contracts are defined as contracts created to advertise or solicit criminal services. These contracts can manifest in various forms, such as pyramid schemes or more dangerous real-world crimes. An example provided is the hypothetical creation of a contract aimed at stealing the Koh-i diamond, which requires the criminal to submit a unique calling card description known only to them. The contract then verifies the completion of the crime by correlating news reports with the description, showcasing the intersection of AI and blockchain technology.
The Role of AI and Oracles in Smart Contracts
To enable rogue smart contracts to access tools like natural language processing and large language models (LLMs), an Oracle system is essential. This system allows smart contracts to effectively communicate with AI technologies by running them off-chain, which is necessary due to the limitations of on-chain processing. By using oracles, smart contracts can gain insights from AI processing, enhancing their functionality and enabling them to execute tasks that rely on external data sources. This convergence highlights the potential and importance of integrating AI capabilities with blockchain applications in smart contracts.
Potential Threats and Safety Measures
The discussion emphasizes the need for caution in engineering systems that combine AI and blockchain technology to prevent malicious uses. Various methods are suggested for managing the integrity and safety of these systems, including employing security councils and techniques developed for AI safety. Additionally, the adoption of pinned models—where the specifications of AI models used in smart contracts are made fully transparent—can be beneficial. This approach aids in preventing unexpected behavior in models, lending an extra layer of safety in decentralized environments.
The Future Role of DAOs and Their Measurement
Decentralized Autonomous Organizations (DAOs) are currently undergoing an experimental phase that seeks to define their success metrics effectively. Traditional measurements such as token distribution fall short of capturing true decentralization, which includes aspects like alignment and diversity of opinions within the community. New metrics, like voting block entropy, have been proposed to measure decentralization more accurately by assessing the distribution of voting blocs instead of individual addresses. This innovative approach aims to advance the understanding of DAOs and their performance as learning systems.
Leveraging Trusted Execution Environments
Trusted Execution Environments (TEEs) are viewed as a promising technology for enhancing security in blockchain applications, including smart contracts and DAOs. They enable confidential transactions and operations, allowing sensitive processes to occur without exposing data on-chain. Despite their vulnerabilities, there is potential for TEE-based systems to offer improved transaction ordering and execution, especially within DeFi frameworks. As research continues, the integration of TEEs with existing blockchain infrastructure may revolutionize the way decentralized systems operate and interact with external data.
A smart contract’s rigid rule system represents a double-edged sword. ‘The Code is Law’, but what happens when rogue large language models or AI agents bend the Law? Can smart contracts be weaponised to serve a criminal agenda? In his ‘Oracle’ novel, Ari Juels explores this thesis and issues an eery warning regarding the blockchain x AI intersection. In this imminent future, oracles play a crucial role, allowing LLMs to push data on-chain or smart contracts to pull off-chain data.
Topics covered in this episode:
Criminal Smart Contracts
Oracle x LLM convergence
Preventing malicious intents through oracle systems
Gnosis: Gnosis builds decentralized infrastructure for the Ethereum ecosystem, since 2015. This year marks the launch of Gnosis Pay— the world's first Decentralized Payment Network. Get started today at - gnosis.io
Chorus1: Chorus1 is one of the largest node operators worldwide, supporting more than 100,000 delegators, across 45 networks. The recently launched OPUS allows staking up to 8,000 ETH in a single transaction. Enjoy the highest yields and institutional grade security at - chorus.one
This episode is hosted by Brian Fabian Crain.
Get the Snipd podcast app
Unlock the knowledge in podcasts with the podcast player of the future.
AI-powered podcast player
Listen to all your favourite podcasts with AI-powered features
Discover highlights
Listen to the best highlights from the podcasts you love and dive into the full episode
Save any moment
Hear something you like? Tap your headphones to save it with AI-generated key takeaways
Share & Export
Send highlights to Twitter, WhatsApp or export them to Notion, Readwise & more
AI-powered podcast player
Listen to all your favourite podcasts with AI-powered features
Discover highlights
Listen to the best highlights from the podcasts you love and dive into the full episode