In this week's Self-Publishing News Special, ALLi News Editor Dan Holloway takes a look at the Association of American Publishers' response to AI, and Kindle Unlimited's first price rise for subscribers.
Keep a lookout for the new self-publishing news podcast later this week. Howard and I have been talking about the German study of young people's reading habits. We had a particularly interesting time thinking about what the increasing clamp down on TikTok in the US meant for social media's reading culture and fiction-centred aesthetic groups (of special note because of the platform's prevalence in driving reading in Germany).
Kindle Unlimited is Raising Subscription Price by 20%
Subscription pricing was a hot topic for a while as Netflix introduced its ad-supported service with a reduced price tag. Things seem to have calmed down of late. As does the debate about what the final form of subscription services more generally will be. But Amazon has just added a little spice by increasing their Kindle Unlimited subscription price for the first time since 2014. Subscribers can keep the current $9.99 by buying 6 or 12 month bundles, which will be available until the end of July. But everyone else will pay $11.99 from their next renewal for the service, which gives them access to all the titles enrolled in KDP Select as well as comics and magazine subscriptions. This makes it more expensive than Kobo’s all-you-can-eat subscription, but brings it into line with Scribd. It will be interesting to see how much trickles into the earnings pot.
I'll add as a breaking item here that KDP print has also announced that print costs will be rising from June 20 – more to come next week!
Scribd says no (ish) to AI
Talking of Scribd, the platform has become the latest to push back against AI. Specifically, it has introduced clauses designed to stop large language model AI scraping the site’s work for training purposes. A press announcement says Scribd will not allow anyone to “utilize the company’s data for monetization or to train large-language models without Scribd’s explicit consent.”
That last phrase feels, to me, somewhat flimsy. Scribd goes on to say that it hasn’t granted any such consent. But as cast iron guarantees go, that feels a little tinny. It’s clear that since Apple and Findaway were found out for burying clauses deep in their terms that would allow their works to be trained, platforms that want authors and publishers to continue using them need to take a long hard look at what they will and won’t allow. On the other hand, big data is big money. As I reported 2 weeks ago, OpenAI has just received a funding round of $300m, giving it a value of nearly $30bn. Those are figures the publishing industry can only dream of. If they come knocking with open wallets, then we’ll see what the guarantees are worth.
Publishing Responds to AI
And in the last AI news of the week, an early sign that this may be on the way to being yesterday’s news is that the Association of American Publishers has finally laid out a position.
To be slightly fairer, Maria Pallante’s talk at the AAP’s Annual General Meeting lays things out in great detail and raises some fascinating questions. We may answer them in different ways, but having the questions before us is vital. I would highly recommend you read Porter Anderson’s in-depth write up and analysis, but I’ll try to break down the key points here.
The Policy of AI
The specific questions she asked seem obvious. How do we protect what she called the inputs (the works that train AI)? And how do we protect the outputs (the works AI produces)? She noted that it looked for a while as though policy makers would take the latter more seriously than the former, but that the tide was turning. Data mining exceptions were being rowed back, for example. This is a good point to add the latest on what governments are actually doing about AI. The European Union has just started work on a major package of legislative rules to govern the use of AI. I will, of course, give more details as they emerge.
The ethics of AI
When it comes to the ethical questions AI raises, I was pleased to see that she grapples with issues around education. Misinformation is a large and real concern around AI. So ensuring that AI is trained on and spits out facts matters. But that raises the question of what count, for intellectual property considerations, as facts. This is an extension of an older and ongoing debate of course around open access and what that means.
But the question that stands out for me is this one. “Do we as a society want AI-generated works flooding the Internet, potentially depressing the value of human authorship?” She goes on to talk about the importance of transparency around provenance. That is, making it clear which content has a human author and which doesn’t. But it’s this value question that’s the most interesting for me as a philosopher, a futurist, and, let’s be honest, as a writer. I expect my answer would differ from that of many (but not all) writers. But I am glad to see the question asked. It’s the perfect topic for ALLi’s new opinion column!Kindle Unlimited raises subscription price by 20%, and other top #selfpub news stories for #indieauthors, in one quick read, by #ALLi News Editor Dan Holloway @agnieszkasshoes #digitaleconomy #publishingopenup Click To Tweet