a16z: Why Does AI Desperately Need Cryptography?
TechFlow Selected TechFlow Selected
a16z: Why Does AI Desperately Need Cryptography?
Artificial intelligence significantly reduces the cost of scaling operations, yet makes trust harder to establish—blockchain technology can rebuild the trust framework.
Navigating Web3 tides with focused insightsBy: a16z
Translated by: Chopper, Foresight News
AI systems are disrupting the internet—originally designed for human-scale interaction—by driving down the cost of collaboration and transactions to historic lows, while simultaneously generating speech, video, and text that are increasingly indistinguishable from human output. We’re already struggling with CAPTCHAs, and now AI agents are beginning to interact and transact just like humans.
The issue isn’t AI’s existence—it’s the internet’s lack of a native mechanism to distinguish humans from machines while preserving privacy and usability.
This is where blockchain technology comes in. The idea that cryptography can help build better AI systems—and conversely, that AI can empower cryptography—rests on several deep logical foundations. Here, we outline key reasons why AI needs blockchain more than ever before.
Raising the Cost of AI Impersonation
AI can forge voices, facial features, writing styles, video content, and even construct full social personas—all at scale: a single agent can operate thousands of accounts, simulating distinct viewpoints, consumers, or voters, with costs continuing to fall.
Such impersonation isn’t new: any motivated fraudster has long been able to hire voice actors, spoof phone calls, or send phishing texts. What’s changed is the cost: today, launching large-scale fraudulent attacks has become dramatically cheaper.
Meanwhile, most online services operate under the default assumption that “one account equals one real human.” When this premise fails, the entire system collapses. Detection-based countermeasures—like CAPTCHAs—inevitably become obsolete, because AI evolves far faster than detection techniques specifically engineered against it.
So how can blockchain help? Decentralized proof-of-humanity or identity systems let users easily complete a one-time verification while fundamentally preventing sybil attacks (i.e., one person controlling multiple identities). For instance, scanning an iris to obtain a global identity credential may be simple and inexpensive—but obtaining a second such credential would be virtually impossible.
By limiting the number of identity credentials issued and raising attackers’ marginal cost, blockchain makes large-scale AI impersonation infeasible.
AI can fake content, but cryptography prevents it from faking human uniqueness at near-zero cost. Blockchain restores scarcity at the identity layer—raising the marginal cost of impersonation without adding friction for legitimate human use.
Building a Decentralized Proof-of-Humanity System
One way to prove humanity is through digital identity credentials—encompassing all information people use to verify their identity: usernames, personal identification numbers, passwords, third-party attestations (e.g., citizenship, creditworthiness), and other relevant credentials.
What does cryptography add? Decentralization. Any centralized identity system at the heart of the internet becomes a single point of failure. When AI agents act on behalf of humans to transact, communicate, or collaborate, whoever controls identity verification effectively controls participation rights. Centralized issuers can arbitrarily revoke access, charge fees, or even facilitate surveillance.
Decentralization flips this entirely: users—not platform gatekeepers—control their own identity data, making credentials more secure and censorship-resistant.
Unlike traditional identity systems, decentralized proof-of-humanity mechanisms let users self-sovereignly manage and store their identity data, enabling privacy-preserving, fully neutral human verification.
Creating Portable, Universal “Digital Passports” for AI Agents
AI agents aren’t tied to a single platform: one agent can appear across chat apps, email threads, phone calls, browser sessions, and APIs. Yet no reliable mechanism currently exists to confirm that interactions across these diverse contexts originate from the same AI agent—with identical state, capabilities, and explicit authorization from its “owner.”
Moreover, binding an AI agent’s identity to only one platform or marketplace renders it unusable elsewhere—fragmenting user experience and making cross-platform adaptation cumbersome and inefficient.
A blockchain-based identity layer enables portable, universal “digital passports” for AI agents. These credentials can encode an agent’s capabilities, permissions, and payment endpoints—and be verified anywhere, significantly increasing the difficulty of forging AI agents. This empowers developers to build more practical AI agents and deliver superior user experiences: agents can operate seamlessly across ecosystems, free from platform lock-in.
Enabling Scalable Payments
As AI agents increasingly conduct transactions on behalf of humans, existing payment systems have become a clear bottleneck. Scaling agent-driven payments requires new infrastructure—specifically, micropayment systems capable of handling high volumes of tiny, multi-source transactions.
Many blockchain-native tools—including rollups, Layer 2 networks, AI-native financial institutions, and financial infrastructure protocols—show strong potential to solve this problem, enabling near-zero-cost transactions and highly granular payment splitting.
Crucially, these blockchain payment infrastructures support machine-scale transactions—including micropayments, high-frequency interactions, and commercial exchanges between agents—none of which legacy financial systems can handle.
- Microtransactions can be split across multiple data providers; automated smart contracts can trigger small payments to all relevant providers in response to a single user interaction;
- Smart contracts enable enforceable, retroactive payments based on completed transactions—compensating parties who provided decision-support information after purchase completion, with full transparency and traceability;
- Blockchain enables complex, programmable payment splits, with rules enforced by code rather than centralized authorities—establishing trustless financial relationships among autonomous agents.
Protecting Privacy and Security Within AI Systems
Many security systems face a paradox: the more data collected to protect users, the easier it becomes for AI to impersonate them.
In this context, privacy protection and security become two sides of the same coin. Our challenge is to design proof-of-humanity systems that are privacy-by-default—hiding sensitive information at every step, ensuring only real humans can disclose precisely the information needed to prove their identity.
Blockchain systems combined with zero-knowledge proofs let users prove specific facts—such as personal ID numbers, passport numbers, or qualification criteria—without revealing underlying raw data (e.g., home address from a driver’s license).
Applications thus gain the necessary identity assurance, while AI systems lose access to the raw data required for impersonation. Privacy protection ceases to be an optional feature—it becomes the core defense against AI impersonation.
Summary
AI drastically lowers the cost of scale—but makes trust harder to establish. Blockchain, meanwhile, rebuilds trust: it raises the cost of impersonation, preserves human-scale interaction patterns, decentralizes identity, embeds privacy by default, and imposes native economic constraints on AI agents.
If we want an internet where AI agents operate smoothly without destroying trust, blockchain isn’t optional—it’s the foundational infrastructure required to build an AI-native internet, and the critical missing piece of today’s web.
Join TechFlow official community to stay tuned
Telegram:https://t.me/TechFlowDaily
X (Twitter):https://x.com/TechFlowPost
X (Twitter) EN:https://x.com/BlockFlow_News
a16z
