The Death of the Identical Paragraph

📊 Full opportunity report: The Death of the Identical Paragraph on ThorstenMeyerAI.com — validation score, market gap, and execution plan.

TL;DR

The longstanding wire service system, built on sharing identical paragraphs, is collapsing due to AI-powered rewriting. Major agencies and publishers are shifting away from syndication toward customized content, raising questions about attribution and funding.

The traditional model of wire services like AP and Reuters, which relied on distributing identical paragraphs to multiple outlets, is rapidly dissolving as AI-driven rewriting technology reduces the cost of producing customized content. This shift challenges the core economic logic of news syndication and raises questions about the future of attribution and funding in journalism.

Historically, wire services pooled the costs of reporting and telegraphing news, allowing newspapers to publish the same information efficiently. This model thrived for over a century, with agencies like AP and Reuters providing international and national news to thousands of outlets. However, recent technological advances—particularly large language models (LLMs)—have drastically lowered the cost of rewriting stories for different audiences, making the traditional syndication of identical paragraphs economically obsolete.

By 2024, the cost of rewriting a 600-word story for multiple sites can be as low as a few cents, significantly cheaper than licensing the same content from a wire service. As a result, publishers and niche outlets increasingly prefer AI-generated, audience-specific rewrites over paying for syndicated content. Major agencies are experiencing declining revenue from traditional sources; for instance, AP’s revenue from U.S. newspapers dropped from around 30% in 2007 to 10% in 2024, prompting diversification into digital and international markets.

This technological shift is exemplified by AI systems that rank, select, and fan out rewrites across hundreds of sites, often rejecting stories that do not meet confidence thresholds. Such systems are already operational in some news organizations, demonstrating the economic viability of replacing the wire’s pooling model with individualized, AI-driven content production.

The Death of the Identical Paragraph — Thorsten Meyer AI
WIRE
● DISPATCH / MAY 2026
THORSTEN MEYER AI · POST-WIRE
POST-WIRE
NEWS / STRUCTURAL ECONOMICS
Essay · News-Industry Structural Economics · 2026-05-15

The Death of the
Identical Paragraph

A 178-year-old labour-pooling arrangement is unwinding underneath the news industry.
Wire copy required everyone to publish the same paragraph for 150 years because no single outlet could afford a foreign correspondent alone. That arithmetic inverted in 2024. AP’s revenue from US newspapers fell from 30% (2007) to 10% (2024). Gannett ended a century-long AP partnership. News Corp signed $250M over five years with OpenAI. The NYT is suing Perplexity over a “skip the click” model and a 96% referral-traffic collapse. The wire is mutating into something else, and who pays for the transition is still being negotiated.
178
Years from AP founding
(1846) to economic inversion
30→10%
AP revenue from US
newspapers, 2007 → 2024
$250M
News Corp–OpenAI
five-year licensing deal
96%
AI-search referral
traffic collapse (TollBit)
AP FOUNDED 1846· REUTERS 1851· HAVAS-REUTERS-WOLFF CARTEL 1865· GANNETT EXITS AP MARCH 2024· NEWS CORP-OPENAI $250M / 5YR· NEWS CORP-META $150M / 3YR· REDDIT-GOOGLE $60M/YR· AP-GOOGLE GEMINI 2025· BARTZ V ANTHROPIC SETTLED $1.5B· MUNICH GEMA RULING NOV 2025· NYT V PERPLEXITY DEC 2025· STEIN 20M LOGS JAN 2026· SUMMARY JUDGEMENT APRIL 2026· AP FOUNDED 1846· REUTERS 1851· HAVAS-REUTERS-WOLFF CARTEL 1865· GANNETT EXITS AP MARCH 2024· NEWS CORP-OPENAI $250M / 5YR· NEWS CORP-META $150M / 3YR· REDDIT-GOOGLE $60M/YR· AP-GOOGLE GEMINI 2025· BARTZ V ANTHROPIC SETTLED $1.5B· MUNICH GEMA RULING NOV 2025· NYT V PERPLEXITY DEC 2025· STEIN 20M LOGS JAN 2026· SUMMARY JUDGEMENT APRIL 2026·
FIG. 01 — AP REVENUE COLLAPSE
The wire’s home audience walked away
AP’s revenue share from US newspapers — the cooperative’s original membership base
2007
~30%
2016
~21%
2024
~10%
AP’s diversification into broadcast (37%), digital ventures (15%), and international (18%) absorbed the gap. In March 2024 Gannett — the largest US newspaper publisher by daily circulation — ended a century-long AP partnership; AP said it was “shocked and disappointed.” Gannett signed with Reuters instead.
FIG. 02 — THE LICENSE STACK
What the AI-publisher deals actually pay
Reported terms from major news-AI licensing agreements signed 2023–2026
PUBLISHER
AI PARTY
REPORTED TERMS
News Corp (WSJ, NY Post, MarketWatch +)
OpenAI
$250M / 5yr
News Corp
Meta
$150M / 3yr
News Corp
Apple
“significant”
Reddit
Google
$60M / yr
Axel Springer (Politico, Insider, Bild)
OpenAI
~$13M / yr
Financial Times
OpenAI
$5–10M / yr
Associated Press
OpenAI
archive · ND
Associated Press
Google · Gemini
terms ND
Agence France-Presse
Mistral · Le Chat
2,300 stories/day · 6 langs
The deals split into training-data licensing (one-shot, archival), display licensing (summaries shown in chat with attribution), and — barely existing yet — raw-feed licensing for downstream rewrite and re-publication. The current dollar volume is roughly $2B cumulative publisher-side. The post-wire economic model needs the third category, and it is not yet contracted.
FIG. 03 — THE COST INVERSION
When rewriting becomes cheaper than not rewriting
Per-story marginal cost, identical-paragraph distribution vs. per-audience rewrite
1846 — 2020
Wire pool
Identical paragraph distributed under N mastheads. Marginal cost of differentiation: a human editor. Marginal cost of identity: telegraph charges divided across subscribers. Identity won, structurally, for 150+ years.
2024 →
Fan-out rewrite
N per-audience rewrites at ~$0.003 each (open-weight, local inference) to ~$0.02 each (cloud-API at the high end). A 50-site fan-out: under one dollar. Differentiation has fallen below the cost of identity.
The wire’s distribution-side logic — pool the cost of the paragraph — is the part that breaks. The reporting-side logic — pool the cost of the bureau in Kyiv — remains intact, and is the part the post-wire model has not yet figured out how to fund.
FIG. 04 — THE LAWSUIT CLUSTER
Where the post-wire rules are actually being written
Active and recently-settled AI copyright cases reshaping news-licensing economics
Dec 2023
NYT v. OpenAI & Microsoft — training-data infringement, “billions” in damages sought · summary judgement scheduled April 2026
In discovery
Sep 2025
Bartz v. Anthropic — authors class action over pirated training data · settled $1.5B, largest US copyright recovery on record
Settled $1.5B
Sep 2025
Penske Media v. Google — first major US publisher suit against Google over AI summaries · ongoing
Active
Nov 2025
GEMA v. OpenAI — Munich Regional Court holds OpenAI liable for German lyrics memorisation · on appeal
Ruled (EU)
Nov 2025
Getty v. Stability AI — UK High Court holds model weights ≠ infringing copies · Getty wins limited trademark on watermarks
Split (UK)
Dec 2025
NYT v. Perplexity — “skip the click” substitution, 175,000 scraping attempts in August 2025 alone, robots.txt ignored
Active
Jan 2026
Stein order, In re OpenAI Copyright Litigation — 20 million de-identified ChatGPT logs ordered into discovery; privacy gambit fails
Ruled (US)
Industry tally: 166 active AI copyright cases as of April 2026, consolidated through MDL or running in parallel. Pattern across rulings: AI companies will pay, eventually, for content used in ways that substitute for the original — rate and mechanism unsettled.
FIG. 05 — THE TRUST PARADOX
Search engines cannot tell good fan-out from bad
Per-site rewrite at scale: structurally what Google claims to want, indistinguishable from what Google is now penalising
17%
Of top-20 Google search
results AI-generated, Sept 2025
50% / 12%
Of new web content AI / share
reaching Google results
45%
Low-value sites cleared by
March 2024 Helpful Content Update
~96%
Referral-traffic drop from
AI search vs. classic search (TollBit)
December 2025 Helpful Content Update reportedly targets “competent but generic” content — pages indistinguishable from fifty others. The signal that separates legitimate per-audience rewrite from undifferentiated AI churn is attribution: a machine-readable, persistent link back to the originating reporter. Whether that link holds is the load-bearing question of the post-wire ecosystem.
Five New York papers founded the AP cooperative in 1846 because no single one of them could afford a correspondent in the field — but five sharing the telegraph bill could. That arithmetic is what has changed.
Thorsten Meyer · The Death of the Identical Paragraph

Why the End of the Wire Changes News Economics

This development signifies a fundamental transformation in how news content is produced and distributed. The decline of the wire service model threatens the traditional funding mechanisms for international and investigative journalism, which relied on shared costs and attribution. As AI reduces the need for syndication, questions about who funds journalism and how attribution is maintained become urgent. The shift could lead to more tailored content but also risks fragmenting the shared information ecosystem that has underpinned global reporting for over a century.

MixPad Free Multitrack Recording Studio and Music Mixing Software [Download]

MixPad Free Multitrack Recording Studio and Music Mixing Software [Download]

Create a mix using audio, music and voice tracks and recordings.

As an affiliate, we earn on qualifying purchases.

As an affiliate, we earn on qualifying purchases.

Historical Role of Wire Services and Recent Disruptions

Wire services like AP and Reuters originated in the 19th century to pool reporting costs, enabling newspapers to access international news without establishing their own bureaus. This cooperative model thrived through the 20th century, with agencies producing most of the international news consumed worldwide. However, declining revenues from print advertising and circulation, alongside the rise of digital media, have strained this model. Recent deals—such as Gannett ending its AP partnership in 2024 and large tech companies investing heavily in AI—highlight the seismic shifts transforming news distribution.

In 2025, AP signed its first deal with Google to incorporate real-time news into Gemini, while legal disputes over AI scraping and attribution have intensified, exemplified by the New York Times’ complaints against AI companies. The core economic rationale—sharing the cost of identical reporting—no longer applies as AI makes rewriting cheaper than syndicating.

“We are exploring new models for local news delivery as the traditional wire service becomes less relevant.”

— Gannett spokesperson

Amazon

news content customization tools

As an affiliate, we earn on qualifying purchases.

As an affiliate, we earn on qualifying purchases.

Unclear Future of Attribution and Funding Models

It remains unclear how news organizations will sustain investigative and international reporting without the traditional pooling model. The future of attribution—whether AI-generated rewrites will carry the same journalistic credibility—is also uncertain. Additionally, the legal and economic frameworks needed to support this transition are still evolving, with ongoing disputes and negotiations.

Amazon

large language model news generator

As an affiliate, we earn on qualifying purchases.

As an affiliate, we earn on qualifying purchases.

Next Steps in News Production and Industry Adaptation

Expect further adoption of AI rewriting systems across diverse outlets, potentially leading to a fragmentation of shared news content. Industry leaders and regulators are likely to develop new standards for attribution, funding, and attribution to address emerging challenges. Monitoring how these changes impact journalistic integrity and economic sustainability will be critical in the coming months.

Amazon

automated news article rewriting

As an affiliate, we earn on qualifying purchases.

As an affiliate, we earn on qualifying purchases.

Key Questions

Will traditional wire services disappear completely?

While their economic model is collapsing, wire services may continue in niche roles or as specialized providers, but their dominant position is likely to diminish significantly.

How will attribution be handled with AI-generated rewrites?

This is still under discussion; some suggest maintaining original source attribution, while others foresee new standards emerging as AI becomes central to content creation.

What impact will this have on investigative journalism?

The decline of shared reporting costs could threaten the funding for international and investigative journalism, raising concerns about information diversity and accuracy.

Yes, issues around copyright, attribution, and unauthorized scraping are active legal debates, with ongoing disputes shaping future regulations.

Will consumers notice a difference in news quality?

Potentially, as AI-generated content may vary in depth and accuracy; the industry is exploring ways to ensure quality and trustworthiness.

Source: ThorstenMeyerAI.com

You May Also Like

The Stanford AI Index 2026 Audit: Reading the Field’s Annual Report Card With a Critic’s Pen

The Stanford AI Index 2026, released three weeks ago, offers a comprehensive report on AI progress, but its methodology and interpretive claims warrant critical review.

The Co-Founder’s Black Hole — A Structural Read on Jack Clark’s Automated AI R&D Essay

Jack Clark predicts over 60% chance that AI will autonomously build its own successor by 2028, raising urgent policy and safety questions.

Agentic Loop Failure Modes: A Production Taxonomy at the End of Year One

A comprehensive taxonomy of failure modes in production agentic systems after one year of deployment, highlighting key categories and operational implications.

Jack Clark Says It Out Loud — Reading the Co-Founder’s 60%/2028 Estimate on Automated AI R&D

Anthropic’s co-founder Jack Clark publicly estimates a 60% probability that autonomous AI systems capable of self-improvement will emerge by 2028.