Starknet, an Ethereum layer-2 blockchain, is set to introduce artificial intelligence (AI) agents that can autonomously carry out on-chain activities for users. These AI agents, developed by Giza, a company dedicated to bringing AI to blockchains, will optimize yield and reallocate portfolios. Giza recently raised $3 million in a pre-seed round led by CoinFund in July 2023. According to Giza co-founder Renç Korzay, the AI agents will be deployed on Starknet by the end of June, with plans to expand to other chains in the future.
The AI agents will initially focus on yield aggregating, identifying the highest return opportunities and depositing funds into different yield avenues. In subsequent releases, the agents will also serve as robo-advisors, determining the best portfolio allocation based on a user’s risk appetite. During the beta phase, the Giza team will review and approve agent strategies to ensure security. Eventually, the agents will become permissionless through the Giza Protocol, which will provide cryptographic and cryptoeconomic security for a trust-minimized exchange between developers and users.
In addition to yield optimization, the AI agents will provide liquidity on decentralized exchanges, make leveraged price bets, and predict price movements. They will adjust crypto holdings based on user-defined strategies. The AI agents will be deployed in phases, starting with the yield-optimizing agents.
Giza’s AI models are built using the Starknet programming language Cairo and utilize Zero-Knowledge Scalable Transparent Argument of Knowledge and zero-knowledge proofs (ZK-proofs) developed by StarkWare. These ZK-proofs enable information sharing and authentication while preserving user privacy.
According to Giza co-founder Cem Dagdelen, the company envisions AI agents becoming the new application layer for Web3, simplifying smart contract risks and complexities for users. The agents are not limited to crypto and blockchain applications; they can also assist highly regulated sectors by verifying the validity of outputs and reducing oversight costs.
While some experts, like Ava Labs founder Emin Gün Sirer, believe that AI-backed smart contract coding is still several years away, others, like Ethereum co-founder Vitalik Buterin, urge caution when integrating AI into blockchains. Buterin believes that while AI can aid in understanding on-chain activities, it should not be responsible for enforcing smart contract rules.
In conclusion, the introduction of AI agents on Starknet represents a significant step towards automating on-chain activities and optimizing user experiences within the blockchain ecosystem.