Andreessen Horowitz (a16z crypto) recently published an article titled “AI needs crypto — especially now.”
The piece, from the a16z crypto editorial team, argues that as AI systems become increasingly capable of generating indistinguishable content (text, voice, video) and coordinating at scale, they’re straining the trust foundations of the current internet, which was built for humans.
Blockchains and crypto provide essential missing infrastructure to restore trust in an AI-native world. Key reasons outlined why AI needs crypto/blockchains right now include: Raising the cost of impersonation and faking human uniqueness — AI can cheaply generate fake content or accounts en masse, but crypto enables “proof-of-personhood” systems like World ID that create digital scarcity for human identity.
It’s easy for a real person to prove they’re human once, but extremely expensive and difficult for AI to impersonate thousands or millions at scale without detection. No single gatekeeper like a centralized platform can dominate verification or participation, reducing risks of centralized censorship or manipulation in an AI era.
Register for Tekedia Mini-MBA edition 19 (Feb 9 – May 2, 2026).
Register for Tekedia AI in Business Masterclass.
Join Tekedia Capital Syndicate and co-invest in great global startups.
Register for Tekedia AI Lab.
Enabling portable, verifiable identities for AI agents — Agents need “passports” that work across platforms without relying on Big Tech intermediaries. Supporting micropayments and agent-to-agent commerce — Traditional payment rails struggle with high-volume, low-value, automated transactions between AIs.
Crypto rails offer fast, low-fee, programmable payments via stablecoins and smart contracts. Privacy by design with tools like zero-knowledge proofs — Allowing verification without revealing unnecessary data, which is crucial as AI handles more sensitive interactions.
The article emphasizes that if we want AI agents to operate autonomously without eroding internet trust via spam, deepfakes, or unchecked coordination, blockchains aren’t optional—they’re the critical layer for an AI-native internet.
This builds on a16z’s ongoing thesis at the intersection of AI and crypto, including prior discussions in their State of Crypto reports, podcasts, and investments in related areas like decentralized AI infrastructure, proof-of-personhood tech, and agentic systems.
The timing aligns with accelerating AI agent adoption and concerns over deepfakes/synthetic media in 2026.
Proof-of-personhood (PoP) systems are mechanisms designed to digitally verify that an online participant is a unique, real human being — not a bot, AI agent, or multiple fake identities created by the same entity.
This addresses a core problem in digital and decentralized systems: Sybil attacks, where one bad actor floods a network with pseudonymous identities to manipulate voting, governance, rewards, content distribution, or spread misinformation.
The concept draws parallels to blockchain consensus mechanisms like proof-of-work (PoW) or proof-of-stake (PoS), but instead of tying influence to computational power or staked assets, PoP ties it to human uniqueness. Each verified person gets roughly one equal unit of participation power, promoting fairness and resisting centralized control or plutocracy.
As AI advances, it becomes trivial and cheap to generate: fake accounts at scale
realistic deepfakes, synthetic text/voice/video
automated spam, scams, or coordinated influence campaigns.
Traditional checks (CAPTCHAs, email/phone verification) are easily bypassed by AI. PoP raises the bar: it’s easy and low-friction for a real human to prove their uniqueness once, but extremely costly or impossible for AI or bad actors to impersonate thousands/millions of unique humans without detection.
This restores scarcity and trust at the identity layer of the internet.In the crypto and AI intersection as highlighted by firms like a16z crypto, PoP is seen as essential infrastructure for: Preventing bot-driven manipulation in decentralized apps, DAOs, or social networks.
Enabling fair airdrops, governance, or resource distribution.
Supporting AI-agent economies where only human-verified entities get certain privileges.
Creating portable, self-sovereign “proof-of-human” credentials that work across platforms without Big Tech gatekeepers. PoP combines verification of humanness (liveness, not a machine) with uniqueness (one person = one credential), often using privacy-preserving tech so no unnecessary personal data is revealed.
Common approaches include: Biometric-based (most robust today): Use unique physical traits hard for AI to fake or replicate at scale.
The leading example is World ID from Worldcoin / Tools for Humanity: Users visit an Orb device, a spherical iris-scanning hardware.
The Orb captures an iris scan to generate a unique, irreversible hash/code proving humanness and uniqueness (irises are highly distinct, even between identical twins). No raw biometric data is stored centrally; instead, cryptographic commitments go into a Merkle tree.
Users receive a World ID credential stored in their wallet/app. They prove membership (i.e., “I’m a verified unique human”) via zero-knowledge proofs (ZKPs) — cryptography that lets you demonstrate a fact (inclusion in the verified set) without revealing which entry you are or any underlying data.
This creates a privacy-preserving “digital passport for humans” usable anonymously across apps. Non-biometric alternatives (explored in research/projects): Social vouching or in-person gatherings. Behavioral analysis or device attestation. Decentralized challenges combining multiple signals.
These are often less secure against sophisticated attacks but avoid privacy and biometric concerns. Zero-knowledge proofs (ZKPs) — Prove you’re in the “verified humans” set without showing who you are or your biometrics.
Blockchain and decentralized ledgers — Store commitments immutably and credibly neutral way, preventing single points of failure or censorship. On-device processing in advanced designs— Ensures sensitive data never leaves your control.
Biometrics raise concerns about data leaks, coercion, or centralization e.g., proprietary hardware like Orbs. Inclusivity — Access to verification and avoiding exclusion of people without tech. Many systems still rely on trusted hardware or operators.
Projects like Worldcoin’s World ID remain the most prominent and scaled implementation in 2026, but the space evolves rapidly with new crypto-native approaches aiming for fully decentralized, open alternatives.
In short, PoP isn’t about revealing who you are like KYC, but proving that you are one real human — a foundational primitive for trust in an AI-saturated, decentralized future.