The intersection of artificial intelligence and blockchain became a hot trend in 2024.

For years, AI projects in the crypto domain primarily focused on decentralizing the infrastructure required for data input and training models. Such AI projects go back to the 2010s. Historically, projects like Numerai, a blockchain-based hedge fund launched in 2015, focused on outsourcing data to crowd-source investment decisions. BitTensor, launched in 2023, adopted a more decentralized approach to developing advanced AI models, including large language models similar to ChatGPT, utilizing user inputs.

This past year, a shift occurred as AI-driven crypto projects began focusing on developing “agents” — AI-powered bots capable of performing tasks such as crypto trading, investment management and user interactions.

This shift was fueled mainly by the unexpected emergence of a chatbot named Truth Terminal. Developed by researcher Andy Ayrey in March 2024, it gained attention for its witty responses on X and was integrated with a crypto wallet. It gained significant attention in July 2024 when venture capitalist Marc Andreessen donated $50,000 worth of Bitcoin to the bot.

Later, one of Truth Terminal’s responses inspired someone to create the meme coin Goatseus Maximus, or GOAT. The developer airdropped a share of the supply to the bot. Once GOAT’s market cap surged beyond $300 million, it propelled Truth Terminal to millionaire status, making it the first AI chatbot to reach this milestone.

The popularity of Truth Terminal spurred a domino effect, leading to the integration of such AI agents into venture capital DAOs that focus on crypto investments. A notable example is ai16z, a community-run tokenized fund that leverages AI and oversees a development framework called Eliza, allowing the creation of other AI agents.

James Ross, founder of Mode — a Layer 2 network focusing on agent development — notes that “an AI agent on a blockchain is essentially a smart contract that can execute decisions independently based on data, without human approval.” Developers can implement access controls to allow agents to perform transactions under specific conditions, expanding the potential applications of AI in blockchain. As such, each agent can function as a dapp with its utility token.

As AI agents begin to manage cryptocurrency wallets and sign keys, new applications are anticipated, ranging from overseeing or verifying nodes on blockchains to executing trades for others. The aim is to enhance user experiences and automate interactions with blockchain-based assets. “Our focus is on developing functional agents capable of executing automated tasks, akin to conventional software-as-a-service applications,” Ross stated.

Eito Miyamura, a crypto researcher and founder of the AI-powered security network GatlingX, thinks that while many use cases exist for AI agents in blockchain, payments present an ideal product market fit. He highlighted the natural synergy between AI agents and blockchain for applications like crypto payments, where agents can operate without the regulatory constraints faced by traditional payment systems.  “The ideal use case for AI agents is when you want an agent working and making crypto payments, purchases, and financial transactions. Crypto is a really interesting area just to do this,” said Miyamura.

According to James Ross of Mode, unlike conventional Web2 systems, crypto enables trustless and permissionless wallet creation, allowing AI agents to transact freely and add value to sectors like DeFi. Ross cited the development of a yield optimization agent named Giza, which operates on the Mode blockchain to maximize DeFi yields. Users can set up a wallet with Giza, which then evaluates various DeFi protocols. The data is fed to an off-chain machine-learning model that analyzes the information and moves the funds into protocols to earn the best yields.

Although the prospects for such technology are promising, similar bots have been used in the cryptocurrency sector for years, such as monitoring large cryptocurrency wallets and issuing alerts when significant transactions occur. Experts say AI models have a key advantage: the ability to continuously update their models, allowing them to learn and adapt from their actions to make better-informed decisions. Furthermore, unlike traditional software bots, these decisions are made without human approval.

“AI agents allow users to gain personalized insights to help them make informed decisions. Unlike traditional software, they can continuously learn and optimize, which is invaluable for complex tasks,” said Ilman Shazhaev, founder of blockchain gaming project Farcana.

Despite the emergence of new applications, AI agents in cryptocurrency still rely on human intervention. Miyamura of GatlingX pointed out that most AI agents are “semi-autonomous,” meaning they operate under human supervision and are far from achieving complete self-control. “They essentially serve as extensions of the developers,” he remarked.

Similarly, James Ross from Mode acknowledged that blockchain-based agents are not fully autonomous, highlighting the need for human involvement.

Ross explained that verifying the output of AI decisions on-chain requires human oversight until a reliable verification method is developed. “AI agents still rely on human intervention because verifying on-chain AI models is crucial to ensure they function as intended.”

However, he remains optimistic, suggesting that it is only a matter of time before agents gain complete autonomy on blockchain networks.

Disclaimer: The Block is an independent media outlet that delivers news, research, and data. As of November 2023, Foresight Ventures is a majority investor of The Block. Foresight Ventures invests in other companies in the crypto space. Crypto exchange Bitget is an anchor LP for Foresight Ventures. The Block continues to operate independently to deliver objective, impactful, and timely information about the crypto industry. Here are our current financial disclosures.

© 2024 The Block. All Rights Reserved. This article is provided for informational purposes only. It is not offered or intended to be used as legal, tax, investment, financial, or other advice.

免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到support@aicoin.com，本平台相关工作人员将会进行核查。