Your AI Agent is Quietly Changing the Rules of the Internet

比推Published on 2026-03-05Last updated on 2026-03-05

Abstract

AI Agents are rapidly transforming the internet landscape, evolving from experimental tools to essential components in daily operations—managing emails, scheduling meetings, and handling support tickets. By 2025, automated traffic is projected to surpass human activity, accounting for 51% of all web traffic, with AI-driven visits to US retail sites surging by 4,700% year-over-year. However, confidence in fully autonomous agents has declined due to security concerns, as infrastructure struggles to keep pace with their expansion. Key challenges include discoverability—agents must efficiently find machine-readable services amidst web pages designed for humans, prompting a shift from SEO to Agent-Oriented Discoverability (AEO). Identity is critical: agents require cryptographic authentication, delegated authority, and real-world accountability to transact securely, leading to emerging standards like ERC-8004 and protocols such as Visa’s Trusted Agent Protocol. Finally, reputation systems are essential to verify agent performance through methods like trusted execution environments (TEEs), zero-knowledge machine learning (ZKML), and economic security models, enabling portable, auditable records of reliability. Together, discoverability, identity, and reputation form the foundational infrastructure for an agent-driven economy, ensuring agents can operate at scale with trust and autonomy.

Author: Vaidik Mandloi

Original Title: Know Your Agent

Compiled and Arranged by: BitpushNews


The promise that AI Agents will change the internet landscape is gradually becoming a reality. They have moved beyond being experimental tools in chat windows to become an indispensable part of our daily operations—from cleaning up inboxes and scheduling meetings to responding to support tickets. They silently enhance productivity, a change often overlooked.

However, this growth is not merely anecdotal.

By 2025, automated traffic has surpassed human traffic, accounting for 51% of total web activity. AI-driven traffic on US retail websites alone has increased by 4,700% year-over-year. AI agents are now operating across internal systems, many with the ability to access data, trigger workflows, and even initiate transactions.

However, confidence in fully autonomous agents has dropped from 43% to 22% within a year, largely due to rising security incidents. Nearly half of enterprises still use shared API keys to authenticate agents, a method never designed for autonomous systems to move value or act independently.

The problem is: agents are scaling faster than the infrastructure designed to govern them.

In response, entirely new protocol stacks are emerging. Stablecoins, card network integrations, and agent-native standards like x402 are enabling machine-initiated transactions. Simultaneously, new identity and verification layers are being developed to help agents identify themselves and operate within structured environments.

But enabling payments is not equivalent to enabling an economy. Because once agents can move value, more fundamental questions arise: How do they discover suitable services in a machine-readable way? How do they prove identity and authorization? How do we verify that the actions they claim to perform actually happened?

This article will explore the infrastructure needed for an agent-driven economy to operate at scale and assess whether these layers are mature enough to support persistent, autonomous participants operating at machine speed.

Agents Cannot Buy What They Cannot See

Before an agent can pay for a service, it must first find that service. This sounds simple but is currently the area of greatest friction.

The internet was built for humans to read pages. When humans search for content, search engines return ranked links. These pages are optimized for persuasion. They are filled with layouts, trackers, ads, navigation bars, and stylistic elements that make sense to humans but are mostly "noise" to machines.

When an agent requests the same page, it receives raw HTML. A typical blog post or product page in this form might require about 16,000 tokens. When converted to a clean Markdown file, the token count drops to about 3,000. This means the model must process 80% less content. For a single request, this difference might be negligible. But when an agent makes thousands of such requests across multiple services, excessive processing compounds into latency, cost, and higher inference complexity.

@Cloudflare

Agents end up spending significant computational effort stripping away interface elements before they can access the core information needed to take action. This effort does not improve output quality; it merely compensates for a web never designed for them.

As agent-driven traffic grows, this inefficiency becomes more apparent. AI-driven crawling of retail and software websites has increased significantly over the past year and now constitutes the majority of total web activity. Meanwhile, about 79% of major news and content websites block at least one AI crawler. From their perspective, this reaction is understandable. Agents extract content without interacting with ads, subscriptions, or traditional conversion funnels. Blocking them is to protect revenue.

The problem is that the web has no reliable way to distinguish between malicious scrapers and legitimate procurement agents. Both appear as automated traffic, both originate from cloud infrastructure. To the system, they look identical.

The deeper issue is that agents are not trying to "consume" pages; they are trying to discover possibilities for action.

When a human searches for "flights under $500," a list of ranked links is sufficient. A person can compare options and make a decision. When an agent receives the same instruction, it needs something completely different. It needs to know which services accept booking requests, what input format is required, how prices are calculated, and whether payment can be settled programmatically. Very few services clearly publish this information.

@TowardsAI

This is why the conversation is shifting from Search Engine Optimization (SEO) to Agent-Oriented Discoverability, often called AEO. If the end-user is an agent, ranking on a search page becomes less important. What matters is whether a service can describe its capabilities in a way that an agent can interpret without guessing. If not, it risks becoming "invisible" in a growing share of economic activity.

Agents Need Identity

@Hackernoon

Once an agent can discover services and initiate transactions, the next major problem is letting the system on the other end know who it is dealing with. In other words: identity.

Today's financial systems run on far more machine identities than human ones. In finance, the ratio of non-human to human identities is approximately 96 to 1. APIs, service accounts, automated scripts, and internal agents dominate institutional infrastructure. Most of them were never designed to have discretion over capital. They execute predefined instructions; they cannot negotiate, choose vendors, or initiate payments on open networks.

Autonomous agents change this boundary. If an agent can directly move stablecoins or trigger a checkout process without manual confirmation, the core question shifts from "Can it pay?" to "Who authorized it to pay?"

This is where identity becomes fundamental, giving rise to the concept of "Know Your Agent" (KYA).

Just as financial institutions verify clients before allowing them to transact, services interacting with autonomous agents must verify three things before granting access to capital or sensitive operations:

  1. Cryptographic Authenticity: Does this agent actually control the keys it claims to use?

  2. Delegated Authority: Who granted this agent permission, and what are its limits?

  3. Real-World Affiliation: Is this agent linked to a legally accountable entity?

These checks together form the identity stack:

  • The base layer is cryptographic key generation and signing. Standards like ERC-8004 attempt to formalize how agents can anchor identity in a verifiable on-chain registry.

  • The middle layer is the identity provider layer. This binds keys to real-world entities like registered companies, financial institutions, or verified individuals. Without this binding, a signature only proves control, not accountability.

  • The edge layer is the verification infrastructure. Payment processors, CDNs, or application servers verify signatures in real-time, check associated credentials, and enforce permission boundaries. Visa's Trusted Agent Protocol is an example for permitted commerce, allowing merchants to verify an agent is authorized to transact on behalf of a specific user. Stripe's Agent Commerce Protocol (ACP) is pushing similar checks into programmable checkout and stablecoin flows.

Meanwhile, the Universal Commerce Protocol (UCP), led by Google and Shopify, allows merchants to publish "capability manifests" that agents can discover and negotiate with. It acts as an orchestration layer and is expected to integrate with Google Search and Gemini.

@FintechBrainfood

An important nuance is that permissionless and permitted systems will coexist.

On public blockchains, agents can transact without centralized gatekeepers. This increases speed and composability but also intensifies compliance pressure. Stripe's acquisition of Bridge highlights this tension. Stablecoins enable instant cross-border transfers, but compliance obligations don't disappear just because settlement happens on-chain.

This tension inevitably draws regulators in. Once autonomous agents can initiate financial transactions and interact with markets without direct human supervision, questions of accountability become unavoidable. The financial system cannot allow capital to flow through unidentified or unauthorized actors, even if those actors are pieces of software.

Regulatory frameworks are already being adopted. The Colorado AI Act, effective February 1, 2026, introduces accountability requirements for high-risk automated systems, with similar legislation advancing globally. As agents begin executing financial decisions at scale, identity will cease to be optional. If discoverability makes agents visible, identity is the credential that makes them recognized.

Verifying Agent Execution and Reputation

Once agents start performing tasks involving money, contracts, or sensitive information, merely having an identity might not be enough. A verified agent can still hallucinate, misrepresent its work, leak information, or underperform.

Thus, the most critical question becomes: Can it be proven that the agent actually did the work it claims?

If an agent states it analyzed 1,000 documents, detected fraud patterns, or executed a trading strategy, there must be a way to verify that this computation indeed occurred and that the output was not forged or corrupted. For this, we need a performance layer to enable this.

Currently, there are three approaches to achieve this:

  1. TEEs (Trusted Execution Environments): The first approach relies on attestation through hardware like AWS Nitro and Intel SGX. In this model, the agent runs inside a secure enclave that issues cryptographic certificates confirming specific code executed on specific data and was not tampered with. The overhead is usually small (around 5-10% additional latency), acceptable for financial and enterprise-grade use cases where integrity trumps speed.

  2. ZKML (Zero-Knowledge Machine Learning): The second approach is mathematical. ZKML enables agents to generate cryptographic proofs that an output was produced by a specific model without revealing the model weights or private inputs. Lagrange Labs' DeepProve-1 recently demonstrated full zero-knowledge proofs for GPT-2 inference, 54-158 times faster than previous methods.

  3. Restake Security: The third model enforces correctness through economic means rather than computational ones. Protocols like EigenLayer introduce staking-based security, where verifiers stake capital behind an agent's output. If the output is challenged and proven false, the stake is slashed. The system doesn't prove every computation but makes dishonesty economically irrational.

These mechanisms address the same problem from different angles. However, execution proofs are episodic. They verify a single task, but the market needs something cumulative. This is where reputation becomes critical.

Reputation turns isolated proofs into a long-term performance history. Emerging systems aim to make agent performance portable and cryptographically anchored, rather than relying on platform-specific ratings or opaque internal dashboards.

The Ethereum Attestation Service (EAS) allows users or services to issue signed, on-chain attestations about an agent's behavior. A successful task completion, an accurate prediction, or a compliant transaction can be recorded in a tamper-resistant way and travel with the agent across applications.

@EAS

Competitive benchmarking environments are also forming. Agent Arenas evaluate agents based on standardized tasks and rank them using scoring systems like Elo. Recall Network reported over 110,000 participants generated 5.88 million predictions, creating measurable performance data. As these systems scale, they begin to resemble real rating markets for AI agents.

This allows reputation to be carried across platforms.

In traditional finance, agencies like Moody's rate bonds to signal creditworthiness. The agent economy will need an equivalent layer to rate non-human actors. The market will need to assess whether an agent is reliable enough to delegate capital to, whether its outputs are statistically consistent, and whether its behavior remains stable over time.

Conclusion

As agents begin to wield real authority, the market will need a clear way to measure their reliability. Agents will carry portable performance records based on verified execution and benchmarking, with scores adjusting for quality decay and permissions traceable to clear authorization. Insurers, merchants, and compliance systems will rely on this data to decide which agents can access capital, data, or regulated workflows.

In summary, these layers begin to constitute the infrastructure of the agent economy:

  1. Discoverability: Agents must be able to discover services in a machine-readable way, or they cannot find opportunities.

  2. Identity: Agents must prove who they are and who authorized them, or they cannot enter the system.

  3. Reputation: Agents must establish a verifiable record proving they are trustworthy, thereby earning ongoing economic trust.


Twitter:https://twitter.com/BitpushNewsCN

Bitpush TG Discussion Group:https://t.me/BitPushCommunity

Bitpush TG Subscription: https://t.me/bitpush

Original Link:https://www.bitpush.news/articles/7617176

Related Questions

QWhat is the projected percentage of automated traffic on the internet by 2025, and what does this indicate about the role of AI agents?

ABy 2025, automated traffic is projected to exceed human traffic, accounting for 51% of total internet activity. This indicates that AI agents are becoming a dominant force, moving beyond experimental tools to essential components in daily operations, silently enhancing productivity in areas like inbox management, meeting scheduling, and support ticket responses.

QWhat are the three core components of the 'identity stack' required for AI agents to operate securely in economic transactions?

AThe three core components of the identity stack are: 1) Cryptographic authenticity (verifying the agent controls the keys it claims to use), 2) Delegated authority (identifying who granted the agent permission and its limits), and 3) Real-world linkage (connecting the agent to a legally accountable entity).

QWhy is 'Agent-Oriented Discoverability' (AEO) becoming more important than traditional SEO for services interacting with AI agents?

AAgent-Oriented Discoverability (AEO) is becoming more critical than SEO because AI agents need machine-readable descriptions of services' capabilities, input formats, pricing, and programmable payment options—not human-optimized web pages with layouts and ads. Without AEO, services risk becoming 'invisible' to agents, missing out on economic opportunities as automated traffic grows.

QWhat are the three methods mentioned for verifying an AI agent's execution and performance to ensure trustworthiness?

AThe three methods for verifying AI agent execution are: 1) Trusted Execution Environments (TEEs) using hardware like AWS Nitro for encrypted certification, 2) Zero-Knowledge Machine Learning (ZKML) for cryptographic proofs of model output without revealing data, and 3) Restake Security, which uses economic incentives (e.g., staking capital) to penalize dishonest behavior.

QHow does reputation infrastructure for AI agents, such as Ethereum Attestation Service (EAS), contribute to the agent economy?

AReputation infrastructure like Ethereum Attestation Service (EAS) allows signed, on-chain attestations of agent behavior (e.g., successful tasks or accurate predictions), creating a portable, tamper-proof performance history. This enables portable reputations across platforms, helping markets assess reliability for granting access to capital, data, or regulated workflows, similar to credit ratings in traditional finance.

Related Reads

Sequoia Dialogue with Jensen Huang: Computing Model Undergoes a 60-Year Transformation; You Won't Be Replaced by AI, But You Will Be Dimensionality-Reduced by 'Those Who Master AI'

NVIDIA founder and CEO Jensen Huang, in a conversation with Sequoia Capital's Konstantine Buhler, argues that we are witnessing the most significant computing shift in 60 years—from retrieval-based to generative computing. Instead of just storing and retrieving data, future systems will generate highly personalized content (text, images, video) on demand, powered by massive "AI factories." Huang envisions a global "intelligence network" that will envelop the planet, following the historical patterns of energy and communication grids. He outlines a five-layer investment framework: 1) Energy, 2) Chips/Computers, 3) Infrastructure (data centers), 4) AI Models, and 5) Applications. He predicts this ecosystem will reach a scale of $20 trillion annually. Crucially, Huang pushes back against fears of AI-driven job loss. He distinguishes between specific "tasks" (e.g., typing, analyzing images) and overall "jobs" (e.g., CEO, radiologist). While AI automates tasks, it increases efficiency and demand for the higher-value problem-solving aspects of professions, thus creating more jobs and "up-leveling" careers. The real risk, he asserts, is not being replaced by AI, but being outperformed by someone who effectively leverages it. He urges everyone to embrace AI as a tool for augmented capability and innovation.

marsbit24m ago

Sequoia Dialogue with Jensen Huang: Computing Model Undergoes a 60-Year Transformation; You Won't Be Replaced by AI, But You Will Be Dimensionality-Reduced by 'Those Who Master AI'

marsbit24m ago

"I Don't Need a Better Model Anymore": A Panorama of AI Users Under a Reddit Hot Post

Titled "I Don't Need a Better Model Anymore": AI User Reactions on Reddit Anthropic recently released Claude Fable 5, its first publicly available 'Mythos'-tier model, achieving 80.3% on the SWE-Bench Pro benchmark and significantly outperforming its predecessor and competitors. However, a viral Reddit post titled "Claude Fable made me realize I don't need better models anymore" highlighted a growing user sentiment of "good enough." Top comments expressed "model fatigue," with users stating that earlier models like Opus 4.5/4.8 already sufficed for their workflows. High cost was a key concern, as Fable 5's API is nearly twice the price of Opus 4.8, with users questioning the return on investment and suggesting the field has hit a plateau. The most frequent complaint targeted Fable 5's stringent safety filters. Designed to intercept high-risk requests (e.g., cybersecurity), the system was perceived as overly conservative. Users reported frequent rejections for routine security-related tasks, leading to automatic fallbacks to the older Opus model. Paying users were particularly frustrated, feeling they paid a premium for a less usable product. Dissenting voices came from users with heavy, complex tasks. For workloads like high-energy physics simulations with thousands of code lines, Fable 5's improved long-context understanding and error detection represented a significant, worthwhile leap—described as moving from a "college player to an NBA starter." The debate underscores a divergence between benchmark performance and practical utility. For most users, current models meet their needs, making further advances relevant only for extreme use-cases. The discussion also raised concerns about a potential "Public AI Freeze," where the most powerful models (like the restricted Mythos 5) remain exclusive to enterprises and governments, while public offerings stagnate. The launch presents two report cards: one of technical excellence and another of user skepticism. Fable 5's ultimate reception may depend on Anthropic's ability to refine its safety filters and justify its cost for specialized, high-demand users.

marsbit31m ago

"I Don't Need a Better Model Anymore": A Panorama of AI Users Under a Reddit Hot Post

marsbit31m ago

When AI Traffic Surpasses Humans, How Do You Prove You're Human?

With AI-generated web traffic surpassing human activity, websites face a crisis as AI agents bypass ads, avoid clicks, and scrape data without generating revenue. This disrupts the ad-based internet economy, diverting traffic and reducing site visits. In response, sites are blocking AI crawlers and deploying traps like Cloudflare's "honeypot" pages. Traditional CAPTCHAs are now ineffective against advanced AI. The focus has shifted to behavioral biometrics—analyzing unique human patterns such as cursor movement, typing rhythm, and keystroke dynamics. Companies like IBM and BioCatch use this data to distinguish humans from bots, even detecting fraud through behavioral inconsistencies. Two competing approaches aim to verify human identity centrally. Sam Altman’s World (formerly Worldcoin) uses iris scanning to create unique credentials, though it faces privacy concerns and regulatory bans. Alternatively, cryptographic zero-knowledge proofs offer anonymous verification without revealing personal data, championed by Vitalik Buterin to avoid centralized surveillance. However, both systems have flaws. Centralized solutions risk biometric data misuse, while decentralized models may be exploited through identity rental markets in economically unequal regions. Despite challenges, the author favors cryptographic methods for preserving privacy over pervasive behavioral monitoring that permanently captures and controls personal biometric data.

marsbit40m ago

When AI Traffic Surpasses Humans, How Do You Prove You're Human?

marsbit40m ago

2026 Landscape of Decentralized AI: Why is Blockchain the Inevitable "Antidote" for AI?

**The 2026 Landscape of Decentralized AI: Why Blockchain is the "Cure" AI Cannot Ignore** Decentralized AI addresses fundamental bottlenecks of centralized AI: scarce and expensive computational resources, excessive control concentration, unverifiable model outputs, and increasing difficulty in acquiring training data due to privacy and regulation. Blockchain offers a path to make intelligence open, verifiable, and economically accessible. The technical stack comprises three layers: 1. **Applications & Services**: The main crypto use cases are "Agentic Finance" (converting natural language into on-chain actions) and "Agentic Payments" for machine-to-machine commerce. Projects like Giza, Infinity Labs, Coinvest AI, and x402 (handling 173M+ transactions) are key players. 2. **Middleware**: This coordination layer enables agents to discover, identify, and transact. Notable projects include Gokite AI (specialized L1), Virtuals (an OS for the agent economy), and especially Bittensor—a network of specialized subnets forming competitive AI micro-economies. 3. **Infrastructure**: The capital-intensive layer providing raw resources. It includes decentralized compute (Akash, Render, Aethir), verifiable inference (Venice AI, OpenGradient), distributed training (Prime Intellect, Templar AI), decentralized storage (Filecoin, Walrus), and privacy/verification layers (Nillion, Arcium, Phala Network) using technologies like ZKPs, MPC, and TEEs. The outlook for 2026-2027 indicates AI demand outpacing infrastructure, with AI agents as a primary growth engine. Computation is becoming an asset class, with on-chain markets as its financial layer. Tokenomics is emerging as a structural advantage for coordinating capital, compute, and data in decentralized AI networks. While still early—with adoption uneven and revenue often trailing token incentives—projects like Bittensor, NEAR, and Virtuals demonstrate a shift from speculative narrative to a new model for coordinating intelligence.

marsbit43m ago

2026 Landscape of Decentralized AI: Why is Blockchain the Inevitable "Antidote" for AI?

marsbit43m ago

a16z Crypto Partner: Cash Flow is the Moat

Cash Flow as the Moat: A Playbook for Crypto Founders Historically, the most enduring businesses have been built by positioning themselves within the "flow of funds"—facilitating the creation and transfer of value in a network and extracting a portion of it. Cryptocurrency is the first modern technology natively built for this purpose. For startups, failing to architect products and businesses to leverage these principles means missing a major opportunity. Blockchains are inherently network businesses. Each transaction settles on a shared ledger, and every new participant strengthens the underlying network for all. Well-designed network tokens amplify this by aligning users, developers, and validators around growing the network, with value flowing back to contributors in a transparent feedback loop. This model is not new; companies from railroads and Standard Oil to Google, Meta, and AWS have thrived by inserting themselves into critical flows of value (goods, attention, compute). Financial markets make it even clearer: firms like Visa and major market makers generate immense revenue not by predicting markets but by being in the path of transactions. The combination of fund flow and network effects creates one of the most durable business structures. The high margins in traditional finance (payments, custody, lending, FX) represent prime targets. Crypto founders have the opportunity to build the next version—programmable, instant, global, and natively in the flow of funds. The frontier extends beyond finance to areas like computing/GPUs, AI training data, energy, robotics, and space—markets without entrenched intermediaries, ripe for building new, efficient value rails on programmable infrastructure. Founders should ask: Are you in the flow of funds today? Does your revenue scale 10x with the value of activity on your platform? Where in your target market are profit margins highest relative to value created? The opportunity is clear: embed your startup into the new flows of value and let the network effects accumulate.

marsbit45m ago

a16z Crypto Partner: Cash Flow is the Moat

marsbit45m ago

Trading

Spot
Futures

Hot Articles

Discussions

Welcome to the HTX Community. Here, you can stay informed about the latest platform developments and gain access to professional market insights. Users' opinions on the price of AI (AI) are presented below.

活动图片