The internet is entering a new phase. Software is no longer limited to displaying information or hosting conversations; it’s beginning to act on our behalf: filtering, recommending, scheduling, purchasing, and negotiating. The “agentic web” can create an unprecedented era of abundance, but it also raises a simple question with serious implications:
When autonomous systems increasingly mediate our data and attention, are we in control?
In the Age of Information, our attention is a precious resource bought and sold by leviathans who convert our digital labor into profit. Our contributions to the system are not investments. We are not stakeholders in our digital futures. We are not decision makers. We must regain power.
When identity, relationships, and data access are controlled by a handful of platforms, our speech and liberty are limited. The remedy is not to replace one gatekeeper with another, but to build an unbiased, decentralized infrastructure where rules are transparent, contestable, and not subject to unilateral seizure.
Infrastructure determines who holds power. Infrastructure is not the loudest layer of the stack, but the foundation that enables the consequential layers above it. There must be “digital rails” for an open and interoperable web. Infrastructure must support:
- Self-sovereign identity and portable relationships, so people are not trapped by platform lock-in
- Explicit permissioning for access to personal data, including the ability to grant and revoke access per application
- Privacy, consent, and ownership as first-class constraints, not afterthoughts or marketing copy
- Compatibility with open standards, including work to interoperate with open-source standards such as the Model Context Protocol (MCP)
The Road Here
The last era of the web delivered real value: global communication, instantaneous information, frictionless publishing. But it also normalized an unhealthy bargain: “free” services paid for through surveillance, behavioral targeting, and the monetization of attention.
Over time, those incentives produced unfortunate outcomes:
- Algorithms that reward outrage and not understanding
- Ecosystems where misinformation spreads as efficiently as truth
- Architecture optimized for clicks, not utility
- Data extractive models consolidating value to an elite few
- Centralized entities that control identity, distribution, and livelihood on private platforms
Our mission is clear: to promote human flourishing by empowering individuals to have full control over their data and attention through a decentralized, open alternative to Big Tech and Big AI.
Five principles for a people-centered agentic web
The path to an agentic web that serves human interests requires a new set of operating principles for digital infrastructure:
1) Verified Representation
Every person should have the right to a single, high-integrity “guardian agent,” a personalized digital fiduciary that stands between you and the internet, verifies information, shields you from manipulative algorithms, and ensures you, not a corporation, control your data and attention.
We need one trusted agentic relationship–an agent that represents your interests, consent, and boundaries across a complex landscape of systems that increasingly compete for your attention and data. This agent must be verified one-to-one through secure proof-of-personhood and anchored to a decentralized ID under the individual’s control.
2) Undivided Loyalty
A guardian must not harm its person, or allow its person to be manipulated or exploited. Guardian loyalty must be legible and enforceable.
Too many digital products are structured with a split allegiance, serving the business primarily and people secondarily. A guardian cannot be built on that compromise. If it is funded, trained, or governed by code that quietly incentivizes exploitation, it will become another instrument of extraction.
A guardian should be required to act in a person’s interest and disclose conflicts plainly when they arise.
3) Data Sovereignty
Consent should be granular, revocable, and understandable.
Genuine choice requires that individuals retain ultimate control over their data, including the ability to withhold, delegate, or revoke its use at any time. Personal guardian agents, acting as loyal digital fiduciaries, should manage how data is used, shared, and transported across services and ecosystems. These agents should always act under the individual’s authority, and also enable individuals to participate in transparent and fair value exchange for the use of their personal data.
4) Cognitive Integrity
Individuals have the right to inspect every algorithmic “nudge.”
No response should reach an individual without being inspected for deceptive intent, hidden persuasion, or manipulative framing. Each individual deserves a personalized, loyal agent to inspect and enforce their boundaries with Big Tech, allowing them to decide whether to permit or limit the data they create.
5) Distributed Governance
Digital ecosystems should be built on neutral protocols.
A technical claim with moral consequences, when an identity graph lives with a single company, it becomes a lever of social power. When it is built on interoperable standards and distributed governance (whether through emerging protocols such as Frequency.xyz or others), optionality is possible, competition is meaningful, and conditions for freedom of expression are possible.
Distributed governance is not lawless. It is transparent, contestable, and not reactive to profit interests.
In the age of the agentic web, Frequency must be leveraged to preserve agency, enable choice, and shift power from platforms to people.
Decentralized Infrastructure for Delegation and Agency
The Frequency project is committed to providing decentralized infrastructure for core functions, especially delegation of access to personal data, so individuals can grant, limit, and revoke permissions as they choose. Maximizing individual agency is a practical path toward driving human flourishing.
DSNP was designed with this in mind: a human-controlled delegation and revocation on a per-application basis. Frequency will operationalize this foundational protocol.
An Open Door to Contributors
Frequency is open-source. Contributions are encouraged from like-minded individuals who want to build digital infrastructure that protects people’s identity, data, and AI context. The core Frequency chain is developed publicly and released under the Apache 2.0 license.
If you share the view that the agentic web should expand human freedom, we hope you will build with us, review our work, and help hold the line for a more people-centered digital future.
Learn more through our Developer Portal.