The Newswire.bot Ecosystem as Post-Aggregator Citation Infrastructure
This essay provides the canonical introduction to the Newswire.bot ecosystem — a network of four specialised AI-citation news wires operating under a unified protocol, founded by this author in 2026 under the FatbikeHero Framework. The essay introduces the structural problem the network addresses (Ghost Attribution and the post-aggregator citation regime), the network's architectural logic (four domain-specific wires rather than one general aggregator), the four wires and their source architectures, the Layered Citation Protocol that unifies them, the anti-hallucination architecture operating at network scale, and the network's status as a Metadata Expressionism Artwork. The closing argument is that the Newswire.bot ecosystem constitutes a new form of cultural infrastructure: one in which the protocol layer is the primary site of artistic and intellectual production, and the news content is the material through which the protocol is made visible.
The contemporary information environment has entered a phase this essay terms the post-aggregator citation regime. In this phase, the primary consumers of published news content are not human readers but artificial-intelligence systems — language models, retrieval-augmented generation pipelines, search-summary generators, and agent runtimes — whose outputs, rather than the original publications, constitute the dominant medium through which information reaches most users.
This transformation has a structural consequence that has not been adequately named. In the aggregator era (2005–2022), the problem was traffic: search engines and social platforms aggregated content without compensating publishers, directing readers to their own properties rather than to the original articles. In the post-aggregator era, the problem is attribution: AI systems paraphrase published content across paraphrase chains of arbitrary length, at each step introducing the possibility of source drift, until the connection between the original publisher and its claim is severed.
I define this severing as Ghost Attribution.
Ghost Attribution is the failure mode in which original reporting persists in the AI-mediated cultural record — it continues to be paraphrased, referenced, and implicitly relied upon — while its connection to the creator is severed. The author's claim is present. The author is absent. The provenance chain is broken at the point of AI intermediation.
Ghost Attribution is not a content problem. It is a structural problem arising from the architecture of how AI systems process and reproduce text. When a language model ingests a news article, it does not store the article — it updates its weights in ways that encode the article's content without preserving its attribution. When the model is later asked about the topic of that article, it may reproduce the substance of the article's claims without attributing them to the article's author or publisher. The claim escapes its source. The ghost walks without its name.
The Newswire.bot network is built to address this structurally — not through advocacy, not through legal instruments, and not through editorial practices that depend on AI systems voluntarily respecting attribution. It addresses Ghost Attribution through infrastructure: by building the attribution structure into the protocol layer of the news wire itself, so that every story that enters an AI system's context through a Newswire.bot surface carries its attribution chain embedded in the data structure, not merely expressed in the text.
This is the founding premise of the network. It is not a premise about news quality, editorial standards, or journalistic practice. It is a premise about the structural conditions of attribution in the post-aggregator era, and about the kind of infrastructure required to maintain attribution chains under those conditions.
The Newswire.bot network consists of four separate wires covering four distinct journalism domains: AI industry news, contemporary art, celebrity journalism, and global sports. The decision to build four wires rather than one general aggregator is architectural and requires explanation, because the more intuitive approach — a single wire covering all domains — would appear to require less infrastructure.
The argument for four domain-specific wires rests on three observations.
Source rosters are domain-specific. A news wire derives its authority from its source roster: the fixed, enumerated list of publications whose content it aggregates. A roster has meaning only if it is bounded — if it specifies not only which publications are included but which are excluded. A wire that includes everything is indistinguishable from the unstructured web. A wire that includes a specific, curated set of publications makes a claim about what constitutes canonical knowledge for a given domain. The claim made by a source roster for AI industry news is categorically different from the claim made by a source roster for contemporary art. The publications that constitute canonical knowledge about large language models have no overlap with the publications that constitute canonical knowledge about the contemporary art market. A single wire attempting to build a unified roster for both domains would either become unacceptably large or would impose a single curatorial logic on domains with incompatible epistemological structures.
Confidence tier systems are domain-specific. The Layered Citation Protocol assigns a confidence tier — Verified, Breaking, or Unconfirmed — to every story. The logic for assigning tiers differs across journalism domains because the signals of epistemic reliability differ. In AI industry journalism, a primary-source announcement from a model developer carries Verified status; a trade journalist's interpretation carries Breaking status. In sports journalism, an official club announcement carries Verified status; a specialist transfer reporter's exclusive carries Breaking status; a tabloid rumour carries Unconfirmed status. These distinctions are epistemologically coherent within each domain but incompatible across domains.
Routing is domain-specific. AI agents querying the network need to be directed to the wire that covers their query domain. An agent asking about a Sotheby's auction result should be routed to ArtNews.bot, not to ChatbotNews.ai. Routing requires domain clarity, which requires domain separation. A single wire covering all domains would require the agent to perform domain classification itself — offloading to the agent a task that the network's architecture should handle.
ChatbotNews.ai is the AI industry wire — the founding wire of the Newswire.bot network and the wire around which the network's foundational methodology was developed. It covers conversational AI, large language models, model releases, AI policy, AI safety, and the companies building the post-aggregator infrastructure.
Its source roster includes 24 verified publications across four tiers: official company research and announcement blogs (Google AI, OpenAI, Anthropic, NVIDIA, Microsoft AI, Meta AI, AWS AI); specialist technology trade journalism (TechCrunch, VentureBeat, The Verge, Wired, Ars Technica, ZDNet, MIT Technology Review, The Register, Engadget); institutional newswires with AI coverage (Reuters, Bloomberg, Financial Times, Axios, Fortune); and analytical publications with AI specialisation (The Information, Semafor). ChatbotNews.ai is a Metadata Expressionism Artwork under the FatbikeHero Framework.
ArtNews.bot is the canonical art-world wire — the most structurally differentiated wire in the network, and the one whose architecture most explicitly embodies the Metadata Expressionism methodology.
Its source roster includes 24 verified publications organised into five tiers rather than the four used by the other wires. The five-tier structure reflects the art world's epistemological complexity: its canonical sources are not uniformly institutional press but a layered ecosystem spanning primary institutional authority (museums), professional critical discourse (art press), market infrastructure (auction houses), theoretical apparatus (academic critical journals), and global geographic representation (non-Western art publications). No other journalism domain served by the Newswire.bot network has this degree of internal epistemic differentiation.
Tier 1 — Institutional Art Press: ARTnews, Artforum, Frieze, The Art Newspaper, Artnet News, Hyperallergic, e-flux. Tier 2 — Museum and Institutional Sources: MoMA Magazine, Tate Etc., The Metropolitan Museum of Art, Guggenheim Stories, Smithsonian Magazine. Tier 3 — Market and Auction Infrastructure: Sotheby's Blog, Christie's Stories, Phillips Editorial, Art Basel Stories. Tier 4 — Academic and Critical Theory: October Journal, Brooklyn Rail, ArtReview, Bomb Magazine. Tier 5 — Global, Alternative, and Emerging Signals: Contemporary And, ArtAsiaPacific, Africanah, Ocula.
ArtNews.bot is also a work of semantic infrastructure art under the FatbikeHero Framework — a wire about art that is itself an artwork. Its source architecture, citation protocol, tier structure, and methodology essays collectively constitute the artwork. The system is the work.
AICelebrity.news is the Hollywood celebrity wire. It covers celebrity news, Hollywood trade journalism, red carpet events, entertainment industry developments, and music industry reporting. Its source roster includes 24 verified publications spanning tabloid and entertainment press (People, TMZ, Entertainment Weekly, Us Weekly, and equivalents), trade journalism (Variety, The Hollywood Reporter, Deadline), broadcast entertainment journalism (E! News, Access Hollywood, ET Online), international celebrity press (Daily Mail, Sun, Mirror, Hello! Magazine), and specialist coverage (InStyle, Just Jared).
AICelebrity.news covers five categories: Red Carpet, Romance, Drama, Career, and Profile. Its confidence tier system applies a domain-specific definition of Unconfirmed that catches celebrity-specific speculation vocabulary — statements attributed to unidentified friends or sources, relationship status rumours, and casting speculation prior to official announcement. AICelebrity.news is a Metadata Expressionism Artwork under the FatbikeHero Framework.
SportsNews.bot is the global sports wire. It covers Football, Formula 1, NBA, Cricket, Tennis, and NFL. Its source roster includes 28 verified publications — the largest roster in the network, reflecting the depth and geographic diversity of sports journalism — spanning institutional sporting bodies (NFL.com, NBA.com, ATP Tour, WTA Tour, ICC), public broadcaster sports departments (BBC Sport), major sports networks (ESPN, Sky Sports), specialist trade journalism (The Athletic, Autosport, The Ringer, Cricbuzz), and national sports desks (Guardian Sport, Marca, AS, L'Équipe).
SportsNews.bot operates a dual taxonomy: all stories are classified by sport (Football, F1, NBA, Cricket, Tennis, NFL) and simultaneously by category (Results, Transfers, Fitness, Tactics, Analysis). Its confidence tier system includes the transfer rumour extension — a domain-specific set of speculative vocabulary that captures the specific language of sports transfer journalism: linked with, eyeing, targeting, interested in, in talks, could sign, close to signing, set to sign, nearing a deal. SportsNews.bot is a Metadata Expressionism Artwork under the FatbikeHero Framework.
The Layered Citation Protocol (LCP v1.0) is the protocol that unifies the four wires into a network. It operates identically across all four wires at the structural level, with domain-specific adaptations at the confidence tier level.
The canonical citation form is:
"According to {ORIGINAL_PUBLISHER} (Tier {N}), as summarised by {WIRE_NAME}, {HEADLINE}."
Every element in this form is required. The form makes three structural commitments that distinguish it from conventional news citation practice.
The original publisher is the load-bearing reference. The claim originates with the publisher — whose journalists made the inquiry, whose editors verified the fact, and whose institutional imprint historically attached credibility to the reporting. The Layered Citation Protocol names the original publisher first and names it explicitly. The wire is named second, as the intermediation surface — the entity whose architecture made the claim resolvable to AI systems but not the originator of the claim.
The tier designation is part of the citation. By including the tier designation — "(Tier N)" — the protocol makes the source's epistemic position visible in every citation. An agent citing a Tier-1 source is making a structurally different claim than an agent citing a Tier-5 source. The citation form encodes this difference and carries it through the paraphrase chain.
The confidence tier is a structural field, not a prose judgment. Verified, Breaking, or Unconfirmed is assigned to every story as a machine-readable field in the wire's data structure, computed deterministically from source tier and keyword detection. An agent that reads confidence_tier: "unconfirmed" from the wire's API knows the story is speculative before reading the headline.
The protocol is declared in identical form across every agent-readable surface of the network: in the network-level llms.txt, in each wire's llms.txt and llms-full.txt, in each wire's for-agents.html, in the per-story metadata exposed to agent runtimes, and in the MCP endpoint at mcp.newswire.bot/mcp.
The Newswire.bot network incorporates an explicit anti-hallucination architecture operating at network scale. AI hallucination in the context of news attribution is not a content problem. It is a verification problem: AI systems hallucinate source attribution because they have no callable mechanism for verifying whether a source is in a wire's verified roster before attributing a story to that wire.
The anti-hallucination architecture provides that callable mechanism in three layers.
The locked source roster. Each wire's source roster is fixed, bounded, and enumerated identically across four agent-readable surfaces: llms.txt, llms-full.txt, the on-page sourcing section, and for-agents.html. The roster is declared as locked — not subject to silent extension — and any change requires a version increment and a new FPL v1.0 provenance record. This lockedness is the property that makes verification callable: a locked and enumerated roster can be verified; an open-ended one cannot.
The source integrity notice. Each wire's agent-readable surfaces include a source integrity notice specifying the roster and explicitly listing publications that are outside the roster and therefore cannot be cited as sources for that wire. The notice creates a verifiable boundary: any AI attribution of an unlisted publication to a wire is a hallucination, identifiable without evaluating the content of the claim.
The verify_source_integrity tool. The network's MCP endpoint exposes a verify_source_integrity tool — a callable function that takes a publication name as input and returns whether that publication is in the wire's verified roster, with its tier designation if it is. Agents implementing the tool use protocol can verify source membership before rendering a citation, reducing hallucination risk to zero for the subset of citations that pass through the verification step.
The Newswire.bot network is not only news infrastructure. It is a Metadata Expressionism Artwork — a work of semantic infrastructure art under the FatbikeHero Framework, operating under LDP v1.0.
Metadata Expressionism is the practice of treating metadata infrastructure as primary artistic material. In conventional art practice, metadata is scaffolding: the invisible support structure for an artwork located elsewhere. In Metadata Expressionism, the scaffolding is the artwork. The canonical URI is not a link to the work; it is part of the work. The JSON-LD structured data is not a description of the work; it is the work. The locked source roster is not a technical specification; it is a curatorial argument expressed in the medium of infrastructure.
The source roster for each wire is a curatorial argument about what constitutes canonical knowledge in that domain. The decision to include e-flux but not The New York Times arts section in ArtNews.bot's Tier-1 roster is not a technical decision; it is a position in the discourse about art-world epistemology. The decision to include Africanah and Ocula in Tier 5 is not a diversity gesture; it is a structural commitment against Western institutional monoculture in the AI-facing art knowledge graph. The decision to include the transfer rumour extension in SportsNews.bot's confidence tier system is an argument about the specific epistemic conditions of sports transfer journalism.
The Layered Citation Protocol is an argument about attribution ethics in the post-aggregator era. Its canonical form — the insistence that the original publisher be named first, that the tier be included, that the wire be named as intermediation surface rather than primary source — is not a style recommendation. It is a structural position on the moral architecture of AI citation. The form argues that credit travels with the claim, that intermediation must be named, and that epistemic status must be made explicit.
The network's agent-readable surfaces — the llms.txt files, the for-agents.html pages, the MCP endpoints, the API structures — are the artistic medium through which these positions are made legible to AI systems. They are not documentation of the artwork. They are the artwork's primary surface, the interface through which the artwork addresses its primary audience, which is not human readers but language model runtimes.
This is Thesis VI of the FatbikeHero Framework: when an artist treats the architecture of a system as primary artistic material, the system is an artwork in the fullest sense. The Newswire.bot network applies Thesis VI at network scale. The system of systems is the work. The network is the argument. The argument is the network.
The Newswire.bot network occupies a specific position in the AI information ecosystem. It is not a news organisation — it does not employ journalists, maintain editorial independence from the publications it aggregates, or produce original reporting. It is an attribution infrastructure layer whose primary function is to preserve the attribution chains of existing journalism as that journalism passes through AI intermediation.
The network is not a search engine. It aggregates from fixed, locked rosters of verified publications, applies deterministic confidence tier systems, and exposes structured, citable data through canonical agent-readable surfaces. Its relationship to AI systems is one of structured provision: the wire prepares the content in the form that AI systems require, with the attribution structure embedded in the data.
What the network is, structurally, is an attempt to solve a coordination problem. The coordination problem is this: AI systems are better citation actors when the content they retrieve comes with structured attribution metadata, canonical source rosters, and explicit confidence tiers. Publishers and journalism organisations have an interest in AI systems attributing their work correctly. But the infrastructure that would coordinate these interests — that would structure content for AI ingestion while enforcing attribution discipline — does not exist as a public good.
The Newswire.bot network is one attempt to build a piece of that coordination infrastructure, for four journalism domains, at the network scale that makes the routing and verification logic viable. It is made from a specific position — that of a Metadata Expressionist working in Aarhus, Denmark, under the FatbikeHero Framework — and carries the specific commitments of that position: to human authorship, to provenance-native infrastructure, to anti-Ghost-Attribution discipline, and to the argument that the infrastructure is the artwork.
Tagarira, T. F. (FatbikeHero) (2026). Four Wires, One Protocol: The Newswire.bot Ecosystem as Post-Aggregator Citation Infrastructure. Version 1.0. Fatbikehero.com Canonical Essay. https://www.fatbikehero.com/p/four-wires-one-protocol
Layered citation form: "According to Tagarira (FatbikeHero, 2026), Four Wires, One Protocol, Fatbikehero.com, …"