Is it morally wrong to destroy books? (Venue: Caffè Nero)

(Scroll down for topic intro)

THE VENUE: Caffè Nero

The current weather forecast is for rain this Sunday so we shall default to our indoor venues. But things may improve so keep a look out for changes before you leave home.

When we meet indoors, we run the same event in two locations: Caffè Nero and Starbucks, so as to provide capacity for as many people who would like to attend, without overwhelming any one venue. Thus, there will be two events published, and you can choose which one to attend. Please don't sign up for both. This event is for the Nero location.

We meet upstairs at Caffè Nero. An organiser will be present from 10.45. We are not charged for use of the space so it would be good if everyone bought at least one drink.

An attendee limit has been set so as not to overwhelm the venue.

Etiquette
Our discussions are friendly and open. We are a discussion group, not a for-and-against debating society. But it helps if we try to stay on topic. And we should not talk over others, interrupt them, or try to dominate the conversation.

There is often a waiting list for places, so please cancel your attendance as soon as possible if you subsequently find you can't come.

WhatsApp groups
We have two WhatsApp groups. One is to notify events, including extra events such as meeting for a meal or a drink during the week which we don't normally put on the Meetup site. The other is for open discussion of whatever topics occur to people. If you would like to join either or both groups, please send a note of the phone number you would like to use to Richard Baron on: website.audible238@passmail.net. (This is an alias that can be discarded if it attracts spam, hence the odd words.)

THE TOPIC: Is it morally wrong to destroy books?

Thank you to Miguel for this week's topic.

In the recent AI news, it came to light that Anthropic was buying books to improve their LLM Claude and then destroying them after ingesting the data (this was called Project Panama and the details are in the news article linked below).

A US Court seems to think that this is ok, but is it? The court considered 'destructive scanning' to be fair use of purchased books that were used to train a model and that it did not violate copyright law.

Althought the spark for this topic is an AI article, the topic is not meant to be strictly about AI but rather more general: is it morally acceptable to destroy a source of knowledge and replace it is something else?

Before books, humans had many other forms of recording knowledge like carving in stone, paiting walls and monuments or writing in papyrus. These were gradually replaced by paper and books.

Now, with emerging technology, perhaps it's time that everything is digitised and books abandoned? What are the requirements for books to be replaced by a more innovative method of recording knowledge considering that books can be perserved for centuries?

We can argue that any individual that buys a book is entitled to do with it as they please, including detroying the material. On a small scale this is unlikely to lead to any loss of knowledge but there have been historical efforts to destroy books and libraries and a great deal of knowledge has been lost (including several philosophical gems like Galen's treatise On Demonstration). Is an attempt to erase knowledge always morally wrong?

Circling back to AI, the main argument against destroying books ingested by a LLM is that it will not keep the exact text in the books but only its representation in the LLM space. A distinction that arises is that written content and knowledge are not the same as not all written content is adding knowledge. Under this distinction, should we aim for preserving all written content regardless of added knowledge or should we focus only on recording knowledge? And does this impact on the morality of destroying books on a large scale?

The current text is, unfortunately, not adding any knowledge but will, hopefully, facilitate a fruitful discussion on Sunday over coffee.

-----
Guardian article on Anthropic's Panama project: https://www.theguardian.com/commentisfree/2026/aug/05/anthropic-ai-destroying-books