On this week's Self-Publishing News podcast, ALLi News Editor Dan Holloway reports on the collapse of Self-Publishing Partnership, a UK-based service provider that ceased trading at the end of June, leaving authors waiting on royalties and raising questions about why no administrators were brought in. The Society of Authors has launched a campaign to help affected members recover what they're owed. Then Dan turns to Anthropic's announcement that it will watermark all text generated by Claude — a move driven in part by the EU's AI Act, and one that could eventually help platforms like Amazon identify AI-generated books.
Listen to the Podcast: Self-Publishing Partnership Stops Trading
Show Notes
How Claude’s text watermark works
Sponsor
Self-Publishing News is proudly sponsored by PublishMe—helping indie authors succeed globally with expert translation, tailored marketing, and publishing support. From first draft to international launch, PublishMe ensures your book reaches readers everywhere. Visit publishme.me.
About the Host
Dan Holloway is a novelist, poet, and spoken word artist. He is the MC of the performance arts show The New Libertines, He competed at the National Poetry Slam final at the Royal Albert Hall. His latest collection, The Transparency of Sutures, is available on Kindle.
Thoughts or further questions on this post or any self-publishing issue?
If you’re an ALLi member, head over to the SelfPubConnect forum for support from our experienced community of indie authors, advisors, and our own ALLi team. Simply create an account (if you haven’t already) to request to join the forum and get going.
Non-members looking for more information can search our extensive archive of blog posts and podcast episodes packed with tips and advice at ALLi's Self-Publishing Advice Center.
Read the Transcript
Dan Holloway: Hello, welcome to the Self-Publishing News. And we start with a story that comes to me courtesy of The Bookseller, and it is another in the fairly lengthy list of unfortunate stories of platforms and companies providing services to indie authors who have gone out of business. It is less high profile than Unbound, which went out of business very messily, leaving very many authors struggling to receive the monies they were owed last year.
Self-Publishing Partnership Stops Trading, Leaving Authors Unpaid
But it is nonetheless something that has left a number of authors out of pocket and wondering when they might see their royalties. So this is Self-Publishing Partnership. It's a UK-based company which provided self-publishing services to enable authors to get their books published and marketed and distributed, primarily in print.
Their distribution partners were Gardners, and what it seems happened is this. So at the end of June, the company's directors sent out an email to their authors stating the usual sort of “with heavy heart” and “in difficult circumstances” wordings that we all know — they will be ceasing to trade.
At that point, there wasn't a timeline given. As I understand it, and again, this information comes courtesy of a very detailed article in The Bookseller, this was then followed up — and The Bookseller indeed followed it up — and only at that point of it being followed up was a move taken to strike the company off from the register of companies at Companies House.
It is very interesting that it has been noted that this was not an attempt to bring in administrators. So normally, the process when a company ceases trading is that administrators will be brought in to try and make sure that everything that was owed to the company was collected, and that then that money that comes in was sent out where it needed to be sent, so that people the company owed money to were paid before the company got wound up.
That doesn't seem to happen, which has raised a lot of questions. The company is not responding to emails, and if you go on their website, you get an error message, so it is very much not clear what is going to happen. What happened when The Bookseller got in touch with them is they noted at least one author got paid some money.
It's a very interesting case study of one of the authors involved. It's not the case that this is a company who was taking a lot of money and then authors weren't really selling books. So it wasn't strictly speaking one of those “you give us lots of money and we give you 10 copies of a book, and that's it” companies.
The author in question had several hundred pounds worth of royalties owing to them each month. They had been paid in the past, and royalties owed were paid during the month of July, but the royalties that are owing for the period after the company ceased trading — no one knows when those are going to be paid.
And I don't think there is any news from any other of the authors involved as to whether they have been paid anything. So one of the continuing issues here is that even though the company had ceased trading, it had nonetheless placed a number of books with Gardners. Gardners had no way of getting royalties to the authors directly.
It appears they have stopped sending royalties to Self-Publishing Partnership and are looking to work to find other ways of making sure that authors who are owed royalties are able to find them, and it is not clear quite what those other pathways are. There is an official campaign by the Society of Authors for its members, so any of its members who are affected can get in touch with them, and they will follow up and try and recover royalties.
I assume that means that they are working with Gardners to do so. But this remains very much a situation in flux, and so my very best wishes to everyone who has been affected by this. As soon as there is news as to how and when people are likely to get their royalties, I will obviously let you know.
But at the moment, the internet is somewhat of a blank. Obviously, the company's page itself is a total blank on this. And there is no evidence that communication has been sent out to their authors since that initial email of the end of June explaining what is happening. So yes. Again, it's one of those unfortunate things that we might not be reliant on publishers, but we are reliant on a number of platforms, and sometimes those platforms go out of business, and we don't have the buffer of publishers or other companies between us and them, and so we are affected directly by what happens to the companies we use.
Also worth noting, and again, I haven't seen anything which suggests anything suspicious about this company, but they weren't on ALLi's register of services. So if you are thinking of using a partner service provider, then do make ALLi's services ratings your first port of call to get the inside info on the people you are thinking of using.
Anthropic Watermarks Claude-Generated Text
And that brings me to watermarks. So watermarks have been in the news. That is the kind of watermark that is traditionally used on paper to verify the authenticity of things like banknotes and stamps and so on, and not the very famous 1990 album by Enya, which is one of the things that I and many of my peers spent all our time as undergraduates listening to.
So watermarks are in the news in relation to AI-generated content because a lot of companies are introducing new ways of watermarking AI-generated content, and a lot of this is driven by the European Union's AI Act, which demands transparency. That is, it demands that it is easy to tell which text or which images have been generated by AI and which haven't.
And the most important development in this is the news from Anthropic that it is introducing a new watermarking system to all text generated by its large language model, Claude. And what I found really interesting is the way that it is suggested that this is going to work. So, Anthropic have a very interesting post on this, and it seems that, unlike sort of DRM, Digital Rights Management, and other watermarking technologies which we might be familiar with in the industry, this isn't metadata, which I guess makes sense 'cause essentially what these models are producing aren't necessarily files, they are just text.
And that text can be taken straight away and pasted into, I don't know, Notepad or equivalent or something that is clearly going to have no metadata with it. So you need to find a way of watermarking that is inherent within the text. That leads to all sorts of questions because we've had a lot of scandals recently around AI detection software and so-called AI tells.
So, it is clearly going to be very controversial if the way you watermark your text is simply by inserting lots of so-called AI tells. It would also be very easy to reverse engineer it out, and this is one of the problems as well when you have something that is entirely embedded within the way you produce your text to prove that it has been generated by AI: how do you stop people working out what it is and then editing the text to edit it out?
The answer is fascinating, and of course, it's one that we're not going to be given, because if we were given the answer, then it would be possible for people to reverse engineer it out. I'm sure that numerous people will already be working on trying to ascertain exactly what it is. But Anthropic assures us of this, that if you were to edit the text, as an editor might edit the text, you will not get rid of watermarked material.
It will still be evident to anyone who has access to Anthropic's own AI detection that you have used Claude to produce your text. If, however, you were to replace everything and totally edit the whole piece from the ground upwards, then they say, well, that probably would get rid of it.
So the answer is that they are doing something to generate the text that lies somewhere in between those two poles. And as I say, exactly what, who knows? It is one step towards trying to make it easier for platforms like Amazon to verify whether text has been AI generated and therefore, to cut down on the number of books competing with us that have been generated by AI.
At the moment, this is just Anthropic. Obviously, the thing is happening that you would imagine is happening, and authors and other people who want to misuse AI to do things that are nefarious are simply closing their Anthropic accounts and turning elsewhere to places that don't do such things at the moment.
So it's very much one of those things that until everyone does it will have a minimal effect. Nonetheless, it's very interesting that at least one company so far thinks it is possible to do this. Whether that is ambitious and wildly over-optimistic, who knows? I'm sure we will have people telling us exactly what they've done very soon.
Nonetheless, it is encouraging to see companies working on this issue. And with that, I hope to bring lots more good news next week and very much look forward to speaking to you then. Thank you very much.
