Investigation · Xavier Vinaixa & Marçal Font Espí
The silent plunder
How a mysterious company buys second-hand books from bookstores across half the world to train AI models — then physically destroys the copies.
This page collects the joint investigation conducted by Xavier Vinaixa (xaviviro) and UB professor, bookseller and poet Marçal Font Espí from the Llibreria Fènix bookshop in Badalona, after spotting anomalous orders from a foreign company systematically buying Catalan non-fiction — often copies that had been sitting in the back room for years.
What looked like an unusual buyer turned out to be an industrial chain: the firms buy the books, ship them to a plant where the spines are cut, the pages are automatically scanned, and the volumes are finally pulped to make paper pulp. All to train artificial intelligence models on human-written text before the web is overrun by AI-generated content — what specialists call the «data wall».
What we uncovered
An international network silently acquires lots of old, low-commercial-value books from second-hand bookshops. They are chosen precisely because they are human texts barely present on the internet — no digital footprint, no editorial scrutiny, no effective copyright protection in practice.
The consequence is not only economic for authors and publishers: it is cultural. Minoritised languages such as Catalan lose written heritage not through human neglect but through algorithmic voracity. Our investigation documents the mechanism and proposes protection and licensing pathways in the Substack trilogy.
Cedulari · Our proposal
The solution we propose: Cedulari
We don't stop at denouncing the problem. For over a year we have been building a constructive answer: Cedulari, the documented memory of Catalan publishing. It is both a database of authors, books and editions —focused on nearly a century of Catalan publishing, from the 1920s to the adoption of the ISBN, a heritage that today is recorded nowhere— and a set of tools that give AI models provenance, traceability and verifiable sources instead of the usual hallucinations.
It was precisely this knowledge —of the second-hand book trade, the print runs, the pseudonyms and the clandestine editions held by only five or six antiquarian booksellers in the country— that let us uncover the silent plunder. Cedulari exists to preserve that heritage and make it available to AI lawfully and traceably. The project, led by Marçal Font, Xavi Vinaixa and Sorensen.ai with advice from the University of Barcelona's literary studies department, is now seeking funding.
Who investigated
01
Xavier Vinaixa Roselló
AI expert, CTO at Sorensen.ai. Pulled the technical thread: identifying the purchase pattern, analysing the logic of the «data wall» and framing it within the wider debate on the real cost of training large language models.
02
Marçal Font Espí
Professor in the Department of Literary Studies at the UB, poet and bookseller, owner of the Llibreria Fènix bookshop in Badalona. Spotted the anomalous orders, documented the requests and contributed first-hand knowledge of the second-hand book trade.
The series: «The silent plunder»
The four instalments of the series are co-authored by Xavier Vinaixa and Marçal Font Espí, published on xaviviro.substack.com, with a companion piece on Marçal Font's Substack (maralfontesp.substack.com):
-
The silent plunder
xaviviro.substack.com →
-
The silent plunder II: the erased confession
xaviviro.substack.com →
-
The silent plunder III: protect and license
xaviviro.substack.com →
-
The silent plunder IV: the architecture of evasion
xaviviro.substack.com · 2026-06-25 →
- Companion piece
And we'll make drums out of the books
maralfontesp.substack.com · 2026-05-18 →
RTVE · TV3 · 3Cat · 2026
On television
The viral 3Cat Info reel
700k+ views3Cat Info's (TV3) report on the literary plunder went viral on Instagram, with over 700,000 views. It features Marçal Font (Llibreria Fènix) and Xavier Vinaixa explaining how some companies buy and destroy books to train artificial intelligence.
«Cafè d'idees» (RTVE) — 15 June 2026
RTVE's «Cafè d'idees», hosted by Gemma Nierga, welcomes Xavier Vinaixa (technical director of Sorensen) and bookseller Marçal Font (Llibreria Fènix in Badalona) to explain how some companies buy books, train artificial intelligence and then destroy them —«they've hit a ceiling, there's no more data»— and to call for European sovereignty in the face of the mass purchase of books to feed AI.
«Tot es mou» — 27 May 2026
TV3's flagship news magazine «Tot es mou» invites Marçal Font (Llibreria Fènix bookshop in Badalona) and Xavier Vinaixa (CTO at Sorensen) on the studio set to explain how they spotted the anomalous purchases, how they pulled the thread to Anthropic and the story originally broken by the Washington Post, and why preserving the physical document is a matter of cultural sovereignty — not just economics. Programme aired in Catalan.
Telenotícies (TV3 newscast) — 31 May 2026
TV3's flagship newscast Telenotícies ran a piece on the case: booksellers denounce the destruction of books to train AI models and call on the Ministry of Culture to step in to protect literary heritage. The report picks up the case uncovered by the investigation. Aired in Catalan.
Press coverage
The investigation has been picked up by television, the press, national radio and international outlets:
- VIA Empresa Opinion ES
Anthropic 451
Opinion piece by Josep M. Ganyet citing the investigation as its starting point.
- elDiario.es Press ES
The mysterious company buying used books to train AI and destroying them
Investigation by Pol Pareja. Quotes Xavier Vinaixa as an AI expert and one of the people who pulled the thread on the case.
- Cadena SER Radio ES
Bread today, hunger tomorrow: a bookseller warns about the AI plan
National radio coverage. Xavier Vinaixa is cited in the audio segment.
- 3Cat / TV3 — «Tot es mou» TV ES
Suspicious second-hand book purchases to train artificial intelligence
Report on «Tot es mou» (TV3) with Marçal Font (Llibreria Fènix) and Xavier Vinaixa (Sorensen) presenting the investigation on the studio set.
- Diario Socialista Press ES
The antiquarian book trade warns about mass purchases of copies to train AI models
Report featuring the warning from UNILIBER (the antiquarian book association). Cites Xavier Vinaixa as an «AI expert who investigated the case» for the «data wall» concept.
- 3CatInfo Press ES
Buying old books to train an AI and destroying them: booksellers' alert over suspicious orders
Report by Toni Noguera Martínez on 3CatInfo. Quotes Xavi Vinaixa (Sorensen.AI) on the «data wall»: the digitized data available for training is estimated to have already hit its ceiling. The report's reel went viral, surpassing 812,000 views.
- Montevideo Portal Press UY
The race to train AI threatens to destroy thousands of old books
Latin American pickup. Cites Xavier Vinaixa as an AI specialist for the concept of the «data wall».
- Estadored.mx Press MX
A Canadian company buys Catalan-language books to train AI and then destroys them
Mexican pickup (Estadored.mx's «Proyecto 84» section) of elDiario.es's investigation into the purchase and destruction of books to train AI. It details Marçal Font's case (Llibreria Fènix, Badalona), the Canadian company Zoom Books, Anthropic's «Project Panama» and the «data wall» concept.
- La Voz del Interior Opinion AR
Artificial intelligence: the new digital extractivism
Opinion piece by Francesc-Xavier Soria Jofra in the Córdoba (Argentina) daily. Extensively cites Xavier Vinaixa and Marçal Font's investigation and reproduces their «Venice phase» passage.
- IADE — Realidad Económica Press AR
The mysterious company buying used books to train AI and destroying them: «It's literary plunder»
The Argentine Institute for Economic Development (IADE / Realidad Económica) republishes Pol Pareja's report (elDiarioAR). Cites Xavier Vinaixa as an AI expert and Marçal Font.
- 3Cat / TV3 — Telenotícies TV ES
«Literary plunder»: booksellers denounce the destruction of books to train AIs
Telenotícies (TV3) news piece on the booksellers' complaint over the destruction of books to train AIs; it picks up the case uncovered by the investigation and calls on the Ministry of Culture to step in to protect literary heritage.
- Telecinco — Informativos Telecinco Press ES
From the bookshop shelf to feeding the algorithm: the controversial purchase of old books to train AI
Report in the Culture section of Informativos Telecinco. Quotes Xavier Vinaixa, technical director of Sorensen.IA: they already hold the information from many books but lack the legal permission to train on it — all they need is the invoice. Also features Miguel Ángel Ortega (UNILIBER) and Marçal Font (UB).
- RTVA — «Les coses grans» Radio AD
Booksellers sound the alarm: US companies buy old books to train AI and destroy them
Panel discussion on «Les coses grans» (Ràdio i Televisió d'Andorra) with Marçal Font (Llibreria Fénix), Toni Noguera (3Cat) and Oliver Vergés (Andorran Publishers' Association). Though not taking part, Xavier Vinaixa is cited several times as the author of the investigation that broke the case.
- Ruido en la Red Social MX
Destroying books to improve AI?
Explainer reel by the Mexican account Ruido en la Red (script by Yunava) on the purchase and destruction of books to train AI. Xavier Vinaixa appears speaking as an AI expert and explaining the case uncovered by the investigation.
- ICI Roussillon — «L'info du Sud» Radio FR
Catalan booksellers worried after mysterious orders bound for the United States
Report by Suzanne Shojaei on «L'info du Sud» (ICI Roussillon, France's public local network). It picks up Catalan second-hand booksellers' worry over suspicious orders bound for the United States and quotes Xavier Vinaixa: these books were probably bought to train artificial intelligence.
- Badalona Comunicació Press ES
Badalona's Fènix bookshop has received several suspicious book orders from a Canadian company
Report by Enric Garcia Torrell in Badalona's public media outlet on the case at the Llibreria Fènix. Quotes Xavi Vinaixa, technical director of Sorensen.AI, warning that the real danger will come when private companies digitize poorly protected books and that information vanishes from accessible databases.
- La Vanguardia Press ES
Destroying books to train artificial intelligence
Report by Lara Gómez Ruiz in the Culture section of La Vanguardia. Extensively quotes Xavier Vinaixa, technical director of Sorensen.ai, and bookseller Marçal Font (Llibreria Fènix) on the investigation published on Substack. Features Vinaixa's proposal to charge these companies for the data and keep the heritage, and presents the Cedulari.cat project.
- LesNews Press CA
Destroying books: a training ground for artificial intelligences
Pickup by the French-language Canadian outlet LesNews. Extensively covers the investigation with Marçal Font («the real challenge isn't buying books but the coming tsunami; the heritage is in danger») and Xavier Vinaixa, featuring his proposal to «charge these companies for the data while preserving our heritage» and to create a consortium and database to protect cultural heritage. Points to Zoom Books and Anthropic's «Project Panama».
- RTVE — «Cafè d'idees» TV ES
How are books destroyed to train AI?
Conversation on RTVE's «Cafè d'idees» with Gemma Nierga. Xavier Vinaixa, technical director of Sorensen.ai, and bookseller Marçal Font (Llibreria Fènix) explain how some companies buy books, train AI and then destroy them («They've hit a ceiling, there's no more data») and call for European sovereignty in the face of the mass purchase of books to feed AI.
- ActuaLitté Press FR
Catalonia: booksellers receive very strange orders for rare books
Report by Nicolas Gary in France's reference outlet for the book world. It picks up Marçal Font's (Llibreria Fènix) alert and the analysis of the suspicious orders published with Xavier Vinaixa, technical director of Sorensen.ai, pointing to the Canadian company Zoom Books.
- RTS — Radio Télévision Suisse Press CH
AI firms buy, digitize and then destroy old books
Report by Swiss public broadcaster RTS (original by Sven Ahnert for SRF, French adaptation by Julien Furrer) following up the investigations by El Diario and Berlin's TAZ. Quotes Xavier Vinaixa, an «AI expert», on the «data wall» concept and the firms' rush to old books once openly accessible knowledge ran dry. Points to the Canadian company Zoom Books.
- Courrier international Press FR
Old books shipped to America are allegedly destroyed after training AIs
Round-up by the French weekly Courrier international, pulling together the investigations by Switzerland's RTS, El Diario (Pol Pareja) and the German daily TAZ on the mass purchase and destruction of books to train AI. It builds on the case uncovered by the investigation and cites Marçal Font (Llibreria Fènix, Badalona), pointing to the Canadian company Zoom Books.
- La Vie Press FR
In Spain, second-hand booksellers plundered by AI
Report by Corine Chabaud in the French weekly La Vie (Groupe Le Monde). Cites both Marçal Font (Llibreria Fènix, Badalona) and Xavier Vinaixa, an «AI specialist computer scientist», who suspect Anthropic is buying second-hand books —through the Canadian intermediary Zoom Books— to train Claude and stave off its «cognitive stagnation». Recalls the «Panama project» revealed by the Washington Post.
- France Culture — «La Revue de presse internationale» Radio FR
Books digitized and then destroyed by AI companies
International press review by Catherine Duthu on France Culture (Radio France). It pulls together the investigations by Switzerland's RTS —spotted by Courrier international—, El Diario (Pol Pareja) and Berlin's daily TAZ on the mass purchase and destruction of books to train AI. Quotes Marçal Font, a Badalona second-hand bookseller, denouncing «a form of literary plunder», and points to the Canadian company Zoom Books and Anthropic's «Panama Project» revealed by the Washington Post.
- El Mundo — Crónica Press ES
Luz's rebellion: the bookseller who refuses to sell books to American AIs to save them from the shredder
Long-form Crónica reportage by Ferrán Barber in El Mundo. It builds on the investigation by Marçal Font (Llibreria Fènix, Badalona) and Xavier Vinaixa Roselló to explain the mass automated purchase of second-hand books from Spanish bookshops to feed AI. It covers Anthropic's «Panama project», the names Zoom Books and PrepFort USA, and Font and Vinaixa's warning about the critical phase to come: when bots go after the printed «intrahistory» not catalogued in the ISBN («We'll end up paying a Silicon Valley company to tell us what our great-grandfather said»).
- El País — Tecnología Press ES
When the internet's knowledge is no longer enough: AI sets its sights on second-hand bookshops
Report by Susana Pérez Soler in El País's Technology section. It reconstructs how the investigation began: the anomalous orders at Llibreria Fènix in Badalona and Marçal Font's call to his friend Xavier Vinaixa, a «researcher and technical director of a technology company», to pull the thread through to ZoomBooks, PrepFort (Illinois) and Anthropic's «Project Panama». Vinaixa explains why destructive guillotining is faster and cheaper than conservation-grade scanning, and why AI companies now need more diverse, better-written data as high-quality public text runs out. With Antonio Ortiz on model collapse and Patrícia Ventura on the privatisation of knowledge.
Also published by: Multimedio El Socavón Noticias de Bariloche Argentinos.es
- Más Información — Código Ciudadano Press MX
They buy second-hand books to train AI
Mexican coverage in the «Código Ciudadano» section, datelined Barcelona. It relays the investigation by Marçal Font (Llibreria Fènix, Badalona) and «researcher and technical director Xavi Vinaixa» into the anomalous orders that exposed destructive scanning of second-hand books. It names ZoomBooks, PrepFort and Anthropic's «Project Panama», framing the story around copyright and the concentration of human knowledge in private corporate repositories.
- IHU — Instituto Humanitas Unisinos Press BR
When the internet's knowledge is no longer enough, AI takes aim at second-hand bookshops
Portuguese translation of Susana Pérez Soler's El País report, published by the Instituto Humanitas Unisinos, the think tank of Universidade do Vale do Rio dos Sinos (São Leopoldo, Brazil). It introduces Xavi Vinaixa as a researcher and technical director of a technology company and carries his explanations of destructive guillotining and the exhaustion of high-quality public text, alongside ZoomBooks, PrepFort and Anthropic's «Panama Project».
- El Colombiano Press CO
Why are millions of second-hand books «disappearing», and what do AI companies have to do with it?
Report by María José Chitiva Londoño in El Colombiano's Tendencias section (Medellín), with its own headline and framing drawn from the El País and Washington Post investigations. It recounts how Marçal Font (Llibreria Fènix) spotted the anomalous order — «They bought four books from me for €35, with €67 of shipping… It made no sense» — and took it to his friend Xavi Vinaixa, director of a technology company, before pulling the thread with other European booksellers on an internet forum. It covers ZoomBooks, the Illinois intermediary PrepFort and Anthropic's «Panama Project», closing on the argument that the goal is not to democratise knowledge but to process it and send the books to the bin.
- Xataka Press ES
AI companies are buying tonnes of old books because they're free of AI slop. Then they destroy them
Report by John Tones in Xataka. It frames the case around human text «free of AI slop» as a scarce resource: it cites Xavier Vinaixa, «the researcher who helped break the case», and analyst Antonio Ortiz agreeing that the industry increasingly needs human text uncontaminated by AI to avoid model collapse. It points to Anthropic's purchase of out-of-print books through intermediaries such as ZoomBooks and ISBNdb and their subsequent destruction, following the 2025 ruling that requires destroying the copies to qualify as «fair use».
- Milenio — Tecnología Press MX
Second-hand books, AI's new target: a bookseller reveals how the scheme to digitise texts works
Report by Lilith Jaime in Milenio's Technology section (Mexico), with its own headline and framing built on El País's investigation. It explains how Marçal Font (Llibreria Fènix, Badalona) spotted the pattern of anomalous orders and pulled the thread with «technology specialist Xavi Vinaixa» through to ZoomBooks (Canada) and the PrepFort logistics centre (Illinois), and how the chain of intermediaries hides the final recipient. It covers Anthropic's «Project Panama» revealed by The Washington Post, destructive scanning and the argument that uncontaminated human content is the antidote to model collapse, closing on the privatisation of knowledge in corporate databases.