a16z: From identity to payments, five reasons why blockchain is a crucial puzzle piece for AI

Author: a16z

Translation: Jiahui, ChainCatcher

AI systems are disrupting an internet originally designed for human scale because they make collaboration, transactions, and the generation of speech, video, and text more affordable than ever, and these generated contents are increasingly difficult to distinguish from human activity. We are already surrounded by CAPTCHAs; now, we are beginning to see intelligent agents interact and transact like humans.

The issue is not the existence of AI; it’s that the internet lacks a native way to differentiate humans from machines while protecting privacy and usability.

This is where blockchain comes into play. The idea of how cryptography can help build better AI systems (and vice versa) may be subtle; therefore, in this article, we summarize several reasons why AI needs blockchain more than ever.

Increasing the Cost of AI Impersonation

AI can mass-produce fake voices, faces, writing styles, videos, and complete social personas: an actor can impersonate thousands of accounts, viewpoints, clients, or voters at decreasing costs.

These impersonation strategies are not new. Any ambitious scammer has long been able to hire voice actors, spoof calls, or send phishing texts. What’s new is the price: scaling these attacks is becoming increasingly affordable.

Meanwhile, most online services assume one account per person. When this assumption breaks down, everything downstream collapses. Detection-based methods (like CAPTCHAs) are bound to fail because AI advances faster than the tests designed to catch it.

So where does blockchain fit in? Decentralized “proof-of-human” or “proof-of-personhood” systems make it easy for a person to participate, but continuously difficult to impersonate many people. While scanning irises and obtaining a World ID may be relatively easy and affordable, acquiring a second one is nearly impossible.

This limits the supply of IDs and increases the marginal cost for attackers, making large-scale impersonation more difficult for AI.

AI can forge content, but cryptography makes it exceedingly difficult to cheaply impersonate human uniqueness. By restoring scarcity at the identity layer, blockchain increases the marginal cost of impersonation without adding friction to normal human activities.

Creating Decentralized Proof-of-Personhood Systems

Proving you are human can be done through digital IDs that contain everything needed to verify identity—usernames, PINs, passwords, third-party attestations (like citizenship or creditworthiness), and other credentials.

What does cryptography add? Decentralization. Any identity system centralized on the internet becomes a single point of failure. When intelligent agents act on behalf of humans—transacting, communicating, coordinating—who controls the identity effectively controls participation. Issuers can revoke access, levy fees, or assist in surveillance.

Decentralization reverses this dynamic: users, not platform gatekeepers, control their identities, making them more secure and censorship-resistant.

Unlike traditional identity systems, decentralized proof-of-humanity mechanisms allow users to control and manage their identities, and verify their human status in a privacy-preserving and trustworthy neutral manner.

Creating Portable, Universal “Passports” for Intelligent Agents

AI agents do not reside in one place. A single agent might appear in chat apps, email exchanges, phone calls, browser sessions, and APIs. However, there is currently no reliable way to know if these different interactions point to the same agent with the same state, capabilities, and authorization provided by its “owner.”

Furthermore, binding an agent’s identity to just one platform or marketplace prevents it from being usable across other products and all other important contexts.

Blockchain-based identity layers enable agents to have portable, universal “passports.” These identities can carry references to capabilities, permissions, and payment endpoints, and can be resolved anywhere, making it harder to forge agents. This also allows builders to create more useful agents and better user experiences: agents can exist across multiple ecosystems without fear of being locked into any particular platform.

Enabling Machine-Scale Payments

As AI agents increasingly transact on behalf of humans, existing payment systems become bottlenecks. Large-scale agent payments will require new infrastructure, such as micro-payment systems capable of handling tiny transactions across multiple sources.

Many existing blockchain tools—rollups and L2 solutions, AI-native financial institutions, and financial infrastructure protocols—show promise in solving this problem, enabling near-zero-cost transactions and more granular payment splitting.

Most importantly, these tracks support machine-scale transactions that traditional finance cannot handle—micro-payments, high-frequency interactions, and business activities between agents.

Nano-payments can be split among multiple data providers, allowing a single user interaction to trigger tiny payments to all contributing sources via automated smart contracts.

Smart contracts enable traceable, executable payments triggered by completed transactions, providing transparent and auditable compensation for information sources that influenced purchase decisions.

Blockchain can distribute complex, programmable payment splits, ensuring revenue is fairly allocated according to code-driven rules rather than centralized decisions, establishing trustless financial relationships among autonomous agents.

Enforcing Privacy in AI Systems

Many security systems face a paradox: the more data they collect to protect users (such as social graphs or biometric data), the easier it becomes for AI to impersonate users.

This is where privacy and security become the same problem. The challenge is to make “proof-of-personhood” systems default to privacy and to obfuscate information at every step, ensuring only humans can generate proof of their humanity.

Combining blockchain-based systems with zero-knowledge proof technology allows users to prove specific facts—PINs, ID numbers, qualification standards (like drinking age at a bar)—without revealing underlying data (such as the address on a driver’s license).

Applications gain the assurances they need, while AI systems are deprived of the raw materials needed for impersonation. Privacy is no longer an overlay but a core defense.

AI offers cheap scalability but makes trust fragile. Blockchain, by raising the cost of impersonation, safeguarding human-scale interactions, enabling decentralized identities, default privacy, and giving intelligent agents native economic constraints, successfully rebuilds trust.

If we desire an internet where AI agents can operate efficiently without eroding trust, blockchain is not optional: it is the key piece that enables a healthy, AI-native internet to function properly.