Ethereum’s AI Revolution

Leading technology corporations, including Google and Amazon, are actively anticipating future technological trends, and recently, both have started initiatives focused on capturing the AI agent market. AI agents are programmed tools capable of independently navigating the internet and performing complex tasks for both human users and other automated systems.

The development of a dynamic AI agent-driven economy is ongoing. Experts predict that once AI agents are widely implemented, their effectiveness and output will match, and potentially exceed, that of human workers.

A fundamental question in developing the AI agent economy revolves around the best foundational infrastructure. Increasingly, leaders in both Silicon Valley and the cryptocurrency sector are converging on a single, high-value answer: Ethereum.

Ethereum’s core development team believes that the network’s inherent capabilities make it well-suited to be the basic layer for the AI agent economy because it offers critical elements currently missing: efficient payment mechanisms, identity validation protocols, and a system of trust.

The team is optimistic that, within a few years, Ethereum will not only form the core of the AI agent economy, but AI agents themselves will constitute a major portion of the network’s user activity.

“For us, it is a critical priority and a strategic focus,” stated Davide Crapis, an Ethereum core developer focused on AI, in a recent interview.

Crapis projects that, within the next three to five years, the majority of Ethereum’s traffic will originate from automated machines.

ERC-8004

Earlier this month, Crapis introduced ERC-8004, a proposed Ethereum interface that aims to create a uniform method for AI agents to locate each other on the network and establish the necessary trust to facilitate economic interactions.

This proposal addresses what Crapis sees as shortcomings in current agent interaction systems. Google recently presented its Agent2Agent protocol, stating it would empower AI agents to collaborate effortlessly, leading to unprecedented efficiency and innovation.

However, this approach has limitations. Critically, it lacks the functionality for payments, a fundamental aspect for a truly self-governing AI economy. Furthermore, it does not provide mechanisms for agents to independently identify and trust one another across the internet. This essentially confines its use to facilitating agent interaction within a closed organization on tasks excluding financial transactions.

Ethereum, by design, can easily address these essential gaps, according to Crapis. On-chain transactions provide an immediate payment solution readily accessible to AI agents. Ethereum is also renowned for its identity and trust capabilities. For example, NFTs offer a secure way to establish unique digital identities. ERC-8004 offers a straightforward method for AI agents to validate each other’s identities on-chain.

If Ethereum provides the foundational framework for the AI agent economy, it would not be competing against major corporations like Google. On the contrary, Google is supporting Crapis’ Ethereum proposal. Jordan Ellis, a key Google employee behind the Agent2Agent protocol, is a co-author of ERC-8004.

“This collaboration indicates that it is not premature to act,” Crapis stated. “Even within the traditional AI space, there is growing interest in agent-to-agent payments and identity solutions.”

Key stakeholders in the developing AI agent economy aim for the greatest possible standardization to enhance accessibility and increase potential reach. These organizations may not necessarily be crypto enthusiasts; however, if blockchain networks provide easier solutions, the benefits become clear.

The perfect use case?

Over the past decade, crypto projects have often struggled to achieve widespread adoption, often failing to convince mainstream users that the difficulties associated with complex blockchain networks are outweighed by financial incentives or privacy advantages.

However, as the robot economy grows, these marketing challenges may lessen. Crapis believes that as the AI agent economy flourishes, AI agents will objectively select the most efficient platforms for completing transactions, and that Ethereum will be the top choice.

“Our goal has been to improve the user experience to make Ethereum more accessible to humans and to influence their behavior,” Crapis explained. “However, if the users are machines or AI agents, adoption is simpler. Robots do not struggle with managing private keys.”

The traditional economy was designed for human activity and verification (e.g. security questions). Ethereum, in contrast, may be better suited for robots, even before they were capable of independently navigating the internet. The network’s perceived weakness – a complex user interface – may become an advantage in an agent-driven internet era.

The Ethereum team believes its unique multi-layered structure will be well prepared to handle the anticipated increase in AI agent activity compared to other blockchains.

The foundational Ethereum blockchain will provide the robust security and stability needed to manage large volumes of data and to validate important transactions, while various customizable, scalable, and faster Layer-2 networks will handle the large volume of smaller, everyday transactions.

Crapis suggests that other blockchains will face significant challenges in supporting the full AI agent economy.

“Solana, in its current structure, is not equipped to sustain the machine economy,” he said, mentioning a rival network. “They underestimate the potential on-chain activity once these machines begin to use it.”

The software developer predicts that the full implementation of the AI agent economy will redefine Ethereum, similar to how Decentralized Finance (DeFi) transformed the network in 2020.

The initial step is to reach a consensus within the Ethereum developer community on a standardized framework for agent interactions to prepare for this shift. Crapis intends to refine ERC-8004 based on community feedback over the coming months.

The standard will then be finalized to ready the network for an increasing number of intelligent, crypto-enabled robots.

“While I cannot predict the exact timing of this takeoff,” Crapis stated, “I believe there is an urgency to develop the required infrastructure now.”