Mike Redwood
Columnist
International Leather Maker
Sitting at our kitchen table sorting through papers while listening to the radio is normally a tranquil moment. But not last week when I heard a second-hand bookshop owner from New Zealand saying that when she received her first bulk order for books, she initially assumed it was a scam.
What was happening was a process which is now termed “destructive scanning”. I think we have all benefited from the mass digitisation of libraries over the last two decades which has given simple access to a huge amount of material that it was difficult or costly to read otherwise. I know from spending time in the National Library in the Netherlands that there is a preservation value in the digitisation of very old books to be read online, allowing them to be handled only by those studying materials, bindings and the details of marginal jottings and other handwritten notes.
Digitising and destroying is not digitising as we knew it
But this was totally different. It was about buying thousand upon thousands of secondhand books regardless of their subject matter and feeding them into an AI system for learning purposes, destroying the books partly before and afterwards, sending the paper for recycling. No one looked to see if there was an unexpected first edition, or if the book was one of few existing copies, if it had significant comments pencilled in the margins by a wise reader, or if it contained details of historical approaches to leathermaking. Of course, all leather books are in the category of “few existing copies”.
In fact, not looking and knowing appears to have been a policy to avoid expensive copyright issues for the AI companies involved, which was initially Anthropic but perhaps others were and still are doing it (I do not know). Anthropic called it Project Panama. Why? So, it could be kept secret for as long as possible; and in this it seems to have been successful.
Some of this information comes from a lengthy copyright court case started in the U.S. in 2024: Bartz vs Anthropic. Unsealed court documents show that Anthropic carried out a massive exercise involving millions of books called “Project Panama”, which they described in an internal memo as an effort to “destructively scan all the books in the world”.
The court decreed that buyers of old books could do what they wanted with them describing it as “fair use’. Many anonymous buyers were involved, so uncovering the truth about how many and what sort of books is hard – deliberately so. Mostly, they were high volume, popular books and in certain cases were genuinely heading for recycling, but several good sources suggest that antiquarian and specialised books were also sought. It is not reassuring.
Inconceivable that amid millions of old books that leather making and history has escaped
I cannot prove that leather books have been involved, but nor can the buyers – Anthropic, their various agents, and others – demonstrate that they have not. They have deliberately not been checking. Given that leather was so important in history, it is hard to imagine that in a million or more books, leather has escaped, and if the programme is allowed to continue as it seems to be doing, more will be lost.
I have issues here. As we look to simplify the process of making old leather, an improved understanding of old processes is essential. Some to be adopted more or less as they were, especially in the craft area where authenticity and history matters, while others will become part of modernised processes suitably adapted for larger scale production of leather free from the difficult chemicals that have become dominant in recent decades.
Keeping our archives and uncovering lost ones is vital for an ancient industry such as leather. Museums like the shoe Museum in Waalwijk, The Netherlands, known as Schoenenkwartier (“Shoe Quarter”), incorporates a research library and knowledge centre while the new Shoemakers Museum in Street, UK, includes a research and study library plus the full footwear maker, Clarks archive. But these great moves are only partially filling gaps left by our major research and educational institutions as they have closed or evolved away from leather. Tannery libraries have largely disappeared with only a few major ones like Friitala (Finland) and Colomer (Spain) being absorbed into local Museums and community archives.
We are already reducing our historic knowledge and archives before we can complete a total inventory. What about ancient literature from Asia and the Arab world that we do not yet know about, where is it hidden? A lot of major leather technology began on the Eurasian plains and around the Caspian Sea and was advancing as various forms of writing were being introduced. They say the origins of writing started with tokens to count animals, grains and evolved into cuneiform to keep better inventories of all manner of stocks which included hides, harnesses and sandals. While tanners were practical folk who passed formulae by demonstration and word of mouth more than on paper, it is hard to imagine that there are not text sources to uncover. The mass destruction of such material in order that some AI bot can churn out language like a Macedonian or Latin scholar is like trying to reverse evolution.
Anyway, the physical book is a cultural artefact; and for most reading online is no substitute, just as typing or dictating is less creative than writing with a pen on paper. The greater durability of paper or parchment over ever changing digital formats (I have floppy discs of various types without any way to read them). The constant updates of software itself routinely make reading older digital material very complex. No wonder they argue that stone tablets are the best storage for data, followed by parchment and then high quality or archival paper: only then followed by digital methods.
Destroying old books is abominable
The first “old” leather book I obtained was being thrown away on the closure of Richardsons tannery in Newcastle. It is “A new and complete treatise on the Arts of Tanning, Currying, and Leather Dressing” by Professor H Dussauce, published in 1865. It is 700 pages of tightly backed formula gathered from all over the world, and I have found it most useful over the last 55 years. The book had been badly treated, not least by being thrown away, with a damaged spine allowing it to fall apart, albeit it was still usable. I found a digital copy freely available online so most of the time I could use that.
After meeting a local bookbinder, Stephen Glendenning of Inventive Bookbinding, at a Leather Conservation Course where I was teaching, I decided to have it properly bound. I provided some of the last pieces of sheepskin leather that I had tanned in India by the Dongola process (not a Dussauce process, as Dongola came from Gloversville a few years later) and my wife gave it to me as a Christmas present two years ago.
It is a magnificent artefact. The first discovery I made was that not only was it stamped with Richardson library stamp, which I had seen before, but it was signed and dated (1865) by David Richardson. He was the family head who decided to rebuild their tannery (I think the sixth over 300 years) on the banks of the River Tyne and opened it in the same year 1865. What is more, in the late 1880s, Henry Richardson Procter worked for three years in the family tannery before moving to Leeds to start the Leather Department and become the leading professor of leather chemistry. He established a laboratory and small research facility in the tannery so this book would have been a major volume in that.
The book was recovered not for display but to be returned for use, and its very detailed Table of Contents and Index means it is much faster and more accurate to find items that you are searching for, and to go on an amazing voyage of discovery when browsing.
The concept that such a magnificent resource and thing of beauty, easy to use, wonderful to show to colleagues is considered with such disregard is abominable. Other major leather books by Procter, Wilson, Waterer and Reed all fall into the same category, as does Beck on gloves and Hoover on the shoe and leather industry. They should not be left to exist as a precious few scattered here and there. Nor should the ones we do not yet know of, with records that might change our view on the material we work on.
One thing is clear and that is while AI might have destroyed two million jobs in mini-drama acting in China, and coding and white-collar areas in the west, leather is too complex to understand and too diverse in its handling for AI to do more than help.
We do not mind help, but we do not accept non-creative destruction. Destroying books like this is not just bad, it is utterly wicked.
mike@internationalleathermaker.com
You can find Dr Mike Redwood across social media at @michaelredwood.
Publication and Copyright of “Redwood Column” remains with the publishers of International Leather Maker. The articles cannot be reproduced in any way without the express permission of the publisher. The views are those of the author and not necessarily those of ILM.
