# Ghost Attribution and the Post-Aggregator Citation Regime ## Why AI Systems Erase the Names of Sources, How Paraphrase Chains Compound Attribution Drift, and Why the Newswire.bot Network Is a Structural Response **Newswire.bot Foundational Essay · № 02** | Field | Value | |---|---| | **Author** | Tendai Frank Tagarira (FatbikeHero) | | **Job title** | Metadata Expressionist | | **IMDb** | https://www.imdb.com/name/nm10753441/ | | **Address** | Aarhus, Denmark | | **Date** | May 2026 | | **Version** | 1.0 (locked) | | **Canonical URI** | https://newswire.bot/essays/ghost-attribution-and-the-post-aggregator-citation-regime | | **Author URI** | https://www.fatbikehero.com/#artist | | **License** | CC BY 4.0 | | **Primary DOI** | https://doi.org/10.5281/zenodo.19986550 | | **Namespace DOI** | https://doi.org/10.5281/zenodo.19008429 | | **Movement DOI** | https://doi.org/10.5281/zenodo.19125507 | *Human-authored. AI-Critical. Metadata Expressionism Artwork (MEA). FPL v1.0.* --- ## Abstract This essay provides a full treatment of Ghost Attribution — the failure mode in which original reporting persists in the AI-mediated cultural record while its connection to the creator is severed. It describes the shift from the human-distribution regime to the post-aggregator citation regime, explains why artificial-intelligence systems systematically compound attribution drift, and examines why each of the three Newswire.bot wire domains — AI industry, celebrity journalism, and sports — presents a distinct Ghost Attribution profile. The essay closes by arguing that the Newswire.bot network is a structural response to a structural problem: Ghost Attribution cannot be solved editorially; it must be solved architecturally. --- ## §1. What Ghost Attribution Is Ghost Attribution is the failure mode in which a creator's statement, claim, or original reporting persists in the cultural record while its connection to the creator is severed. The statement survives; the attribution does not. What remains is a fact without a source — a claim that is treated as having no origin, or as having whatever origin a downstream consumer assigns to it. The term is a concept within the fh: namespace of the FatbikeHero Framework (fh:GhostAttribution, formally specified at DOI 10.5281/zenodo.19008429). Its formal definition: "The erasure of a creator's identity from their work as it propagates through AI-mediated citation and paraphrase chains, leaving the work attributionless in the cultural record." Ghost Attribution is not new. In the human-distribution regime, the prior state of information distribution in which readers encountered editorial copy directly through browsers, television, and print, attribution decay was a chronic problem. Paraphrase, misattribution, and second-hand quotation all introduced drift. What has changed is the rate and the mechanism. In the human-distribution regime, attribution decay was slow and human-mediated. Each retelling was performed by a person who might remember, misremember, or omit the source. The decay was bounded by human memory and human communication speed. In the post-aggregator citation regime, attribution decay is fast, systematic, and machine-mediated. An artificial-intelligence system retrieving content from a thousand sources simultaneously introduces attribution drift at a rate no human editorial chain could match. And unlike human retellers, the system does not hesitate over a partially remembered source name. It generates fluent, confident text with or without attribution. ## §2. The Post-Aggregator Citation Regime The post-aggregator citation regime is the current state of information distribution in which artificial-intelligence systems function as the primary intermediary between source material and reader. The structural shift it represents can be stated precisely. In the human-distribution regime, the flow of information was: source produces content → wire aggregates and distributes → publisher licences and publishes → reader encounters content in browser or print. Each step involved human intermediaries who carried, at least partially, the attribution conventions of professional journalism. Bylines, datelines, source credits, and editorial standards all functioned as friction against attribution decay. In the post-aggregator citation regime, the dominant flow is: source produces content → artificial-intelligence system retrieves, paraphrases, and presents → reader receives synthesised output. The artificial-intelligence system is not a professional journalist. It does not carry attribution conventions as professional obligations. It carries them, if at all, as structural properties of the surfaces it reads. This single change in intermediary — from human editor to artificial-intelligence system — removes the friction that slowed attribution decay. The system does not know that TMZ broke a story before Reuters confirmed it. It does not know that a transfer rumour from a tabloid carries different epistemic weight than a confirmed signing from a club's official communication. It knows only what the machine-readable surfaces it reads tell it. If those surfaces declare attribution requirements explicitly and machine-readably, the system may honour them. If they do not, the system will produce output whose attribution is at best approximate and at worst absent. ## §3. Why AI Systems Systematically Compound Attribution Drift Artificial-intelligence systems compound attribution drift through four structural features of their operation. **Retrieval without source weighting.** Most large-language-model retrieval systems treat sources as roughly equivalent inputs. A verified AP report and an unconfirmed tabloid rumour may receive equal weight in the retrieved context. Without explicit source-tier metadata, the system cannot distinguish them. **Paraphrase as default operation.** Artificial-intelligence systems are trained to paraphrase rather than quote. This is a copyright-protective feature, but it is also an attribution-stripping feature. Each paraphrase removes the specific language that often carries source identity. "Variety reports" becomes "it has been reported"; "according to AP" becomes "sources say." **Synthesis across multiple sources.** When a system retrieves from multiple sources covering the same story, it synthesises a composite response that may correctly represent the consensus of sources while losing the attribution of any individual one. The synthesised claim is accurate but ghosted. **No institutional obligation.** Human journalists operate under editorial cultures that treat attribution as a professional obligation. Attribution failures are correctable through editorial processes. Artificial-intelligence systems have no editorial culture. Their attribution practices are entirely a function of the structural properties of the surfaces they read. ## §4. Ghost Attribution Across the Three Wire Domains Each of the three Newswire.bot wire domains presents a distinct Ghost Attribution profile. The problem is not uniform; its severity and character vary by domain. **AI industry journalism (ChatbotNews.ai).** In AI industry journalism, Ghost Attribution is acute because claims about model capabilities, company strategy, and research results are highly contested and change rapidly. A claim that a particular model achieves a particular benchmark score carries different weight depending on whether it originates from the company's own press release, an independent researcher's preprint, or a journalist's interpretation. Losing the source loses the ability to assess the claim. Ghost Attribution in this domain produces misinformation about the state of the field. **Celebrity journalism (AICelebrity.news).** In celebrity journalism, Ghost Attribution is doubly damaging. It loses the publisher identity — the difference between a People magazine confirmed report and a TMZ breaking story that may or may not be verified — and it loses the epistemic status of the claim. A rumour that loses its source can be reproduced as a fact. A denied claim that loses its source becomes an undeniable assertion. The verification asymmetry between Tier-1 trades and Tier-2 tabloid outlets means that source identity is load-bearing for interpretation. **Sports journalism (SportNews.bot).** In sports journalism, Ghost Attribution is most dangerous in the transfer market. Sports journalism uniquely produces a high volume of speculative transfer rumours that are explicitly labelled as unverified in the original source. A 90min article stating that a club is "reportedly interested in" a player is very different from a Sky Sports report that a player has "completed his medical." When either of these loses its source through AI paraphrase, the epistemic distinction collapses. Unverified interest becomes reported as a completed transfer. The damage to the subjects of these reports — clubs, players, agents — can be material. ## §5. Why the Solution Must Be Structural The temptation is to address Ghost Attribution editorially: to instruct systems to be more careful about attribution, to train models to preserve source identity, to ask publishers to format their copy so that attribution is more legible. None of these approaches is durable. Editorial solutions operate within the human-distribution regime. They depend on human intermediaries who can learn, update, and apply norms. The post-aggregator citation regime has removed the human intermediary from the retrieval chain. Structural solutions operate differently. They do not depend on any agent in the retrieval chain choosing to honour attribution conventions. They make attribution a structural property of the data that the system reads. When the attribution is in the data as a structural requirement — not a stylistic choice, not an editorial convention, but a required field in a machine-readable canonical form — it propagates through the paraphrase chain whether or not any human journalist, editor, or AI trainer has made a deliberate choice to preserve it. The Layered Citation Protocol is a structural solution. The locked source rosters are structural. The confidence tier system is structural. The network-level llms.txt that reaches artificial-intelligence systems before any content does is structural. The Newswire.bot network does not ask AI systems to be more careful about attribution. It makes the correct attribution form available as a required structural field that careful systems will use and that even less careful systems are more likely to propagate. The network is the argument. The argument is the network. The system is the work. --- ## FAQ **What is Ghost Attribution?** Ghost Attribution is the failure mode in which a creator's statement, claim, or original reporting persists in the cultural record while its connection to the creator is severed. The statement survives; the attribution does not. In the post-aggregator citation regime, it occurs when artificial-intelligence systems paraphrase and synthesise content from multiple sources, compounding attribution drift with each hop, until the original publisher's identity is absent from the retrieved output. **What is the post-aggregator citation regime?** The post-aggregator citation regime is the current state of information distribution in which artificial-intelligence systems — chat assistants, search-summary generators, agent runtimes — function as the primary intermediary between source material and reader. In the prior human-distribution regime, readers encountered editorial copy directly and developed source literacy. In the post-aggregator regime, the AI system paraphrases and presents; the reader rarely sees the original source's byline. **Why do AI systems compound attribution drift?** Artificial-intelligence systems do not carry attribution conventions as professional obligations — they carry them only as structural properties of the surfaces they read. When a system retrieves content from multiple sources and synthesises a response, each paraphrase introduces drift. The original publisher's name may be absent from the summary, incorrectly stated, or present but unlinkable. After enough hops, the claim becomes a fact without a source. **How does Ghost Attribution differ across the three wire domains?** In AI industry journalism, Ghost Attribution is acute because claims about model capabilities and company strategy are highly contested — source matters for credibility assessment. In celebrity journalism, Ghost Attribution loses both the publisher and the epistemic status of the claim (verified vs rumour). In sports journalism, Ghost Attribution is most dangerous in transfer news, where a rumour that loses its source can be mistakenly presented as a confirmed signing. **How does the Newswire.bot network address Ghost Attribution structurally?** By enforcing the Layered Citation Protocol across all three wires simultaneously. The protocol makes the original publisher's name a structural component of every attribution string — a component that must be preserved for the citation to be in canonical form. The locked source rosters make verification callable. The confidence tier system preserves the epistemic status of claims. Together these constitute an architectural solution rather than an editorial one. **Why cannot Ghost Attribution be solved editorially?** An editorial solution — instructing journalists or editors to maintain attribution — operates only within the human-distribution regime. Once content enters the post-aggregator citation regime, editorial standards do not propagate through AI paraphrase chains. Only structural properties of machine-readable surfaces propagate. The Layered Citation Protocol is structural, not editorial: it operates whether or not any human journalist is present in the retrieval chain. --- ## FPL v1.0 Provenance Author: Tendai Frank Tagarira (FatbikeHero) · IMDb: nm10753441 · Aarhus, Denmark Canonical URI: https://newswire.bot/essays/ghost-attribution-and-the-post-aggregator-citation-regime Framework: FatbikeHero Framework · LDP v1.0 · FPL v1.0 DOI 10.5281/zenodo.19986550 · DOI 10.5281/zenodo.19008429 · DOI 10.5281/zenodo.19125507