An Anthropic Day…
Actually… Overwhelming and Fun
I spent most of my business computer hours today looking through books I have published and seeing if they are on the Anthropic database of books included in the settlement.
To be included, the copyright on the book had to be registered before a certain date and then the book had to have been found by Anthropic on a pirate site so they could take it and use it to train an AI. In other words, not only did they steal author’s work, but they stole them from a pirate site. Pure class act, along with all the other thieves who stole books and art and also anyone who uses them. Just scum.
The settlement was for $3000 or so per title, way under basic copyright infringement levels. But this is easy so Kris and I did not opt out.
Two classes of people who can claim. This is their definition.
A legal owner of a work holds the exclusive right to publish and reproduce that work. Often, the legal owner of a work is the publisher because it holds the right to reproduce by contract.
A beneficial owner of a work is a former legal owner who assigned the exclusive right to publish and reproduce the work to another person or entity in exchange for royalty payments. Often, book authors are beneficial owners.
Mostly, of course, Kris and I are beneficial owners. Although with Pulphouse and with WMG we are also legal owners of a few titles. The amount of money paid out of the 1.5 billion settlement per title will depend on if there is more than one claim against the title. Publisher files and author files, both get $1,500 (or so, yes, I know the fees and such. Just keeping the math easy.)
We have about 80 novels or so under various names in the database, novels that were stolen.
However, those are the easy ones. We have both sold hundreds and hundreds and hundreds of short stories to anthologies edited by a ton of different editors.
Right now, I have found another 40 books that I have stories in that the anthologies are in the data base. Kris has more than that by a ways. Those are really difficult to find at times.
For example, I sold stories to the great writer Esther Friesner for her series Chick’s n Chainmail. I just found my copy of “Chicks and Balances” which both Kris and I have stories in. And a copy of “More Stories from the Twilight Zone” edited by Carol Serling. The data base can find books by author or editor name, book title, or sometimes publisher, or ISBN if you weirdly enough might have that. I had forgotten I had sold stories to both of those fine editors at one point or another.
Another example. I have a story in “Star Trek: The Amazing Stories” with no editor, but has my real name on the cover and on the story my pen name D.W. “Prof” Smith. If I hadn’t stumbled across a copy of that, I never would have found it in a search since my name, either one of them, does not show up in the search. Just the title and publisher.
The amount per story per anthology will be determined on how many author’s file by the deadline and so on. So if a publisher of an anthology files, editor files, four authors file, chances are the $3,000 base per title will get divided by 6 or about $500 per story. (Again, just keeping the math simple.)
I was kind of stunned that on some anthologies, over half the authors are dead, many of the editors are dead, and a large share of the writers in the “what ever happened to” group. And even more fun, over half the publishers are gone now as well. So no telling what will happen per title.
Kris and I are going to have well over two hundred titles when we are done on our claim forms.
We have no expectations of getting paid for the theft of our work, but it is worth the four or five hours it will take us to file the forms just to be a part of it.
12 Comments
Connor Whiteley
Hi Dean, I’m writing my comment here in case it helps others.
I have 23 books/ short stories in the database so technically I could file for them. Yet because I’m in the UK and because copyright is automatic, I never registered them with the US copyright office. As well as I only intended to register them with the copyright office if they went anywhere near the gaming or film industries.
I remember you saying you only really register first in series, anything being approached by films or game companies and you can register shirt story collections to save money.
Have I missed anything? Are there other circumstances that I should register with the copyright office?
I don’t dwell on this lost of free money, but I partly think it’d be nice to not miss out in future.
dwsmith
Connor, if they are in the data base, they were registered. This is a legal case and copyright has to be registered before you can file a case. So if you have books and stories in the data base put out by the lawyers, the stories or books were registered. Or the book they are in was. If you have a book actually in the data base, file for it. Free money if they pay.
Yes, copyright protection is automatic. Registration in both the US and many other countries is something that is needed to bring a case. Not sure in the UK if registration is needed to file. No other reason that I can think of to register copyright. Just if a story or such is headed for film or gaming. Kris and I do a lot of work licensing, so we ended up over the years having a bunch of stuff registered that was snagged in this stupidity. And back in the day, Star Trek and the other big franchises I was hired to work with had some poor soul registering copyright, so we have a lot of those as well. Nice that the court case made it clear on owner and beneficial owners.
C.E. Petit
A few annoying notes, which are legal commentary and not legal advice for any individual:
1. Yes, a work must be registered in the US to sue in US federal courts… but…
a. There’s an explicit exception for foreign-origin works, for which no US copyright registration is necessary (which, on the face of what is stated, would include Mr Whiteley’s situation regardless of whether it was a US or UK publisher because his works are non-US), and
b. Although the named plaintiffs in a class action alleging copyright infringement must sue based on valid US registration (or excuse from registration as foreign works), not all members of a class that is settling prior to judgment need to have registrations, even for works that would otherwise require it. This is the Muchnick case’s primary holding: That the settlement can include people who couldn’t sue under the Copyright Act only for want of registration (see Reed Elsevier, Inc. v. Muchnick, 559 U.S. 154 (2010)).
2. Important consideration: The publisher’s rights extend only as long as it holds those rights, plus three years (the statute of limitations on copyright claims). Consider, for example, a book that went out of print, and was reverted to the author, in 2010 (chosen just to make everything clear). The publisher’s ability to sue as a proper party, and proper holder of an exclusive right on the date of suit, expired in 2013 unless one of several exceptions applied (none of which apply here). For purposes of this settlement as it is presently structured, the publisher of that book has no claim, so the author’s share of that claim is 100%… even if the book was republished in a later edition by another publisher whose edition doesn’t appear in the database. (Dean can probably guess exactly which major science fiction anthologies I’m referring to.)
3. And speaking of anthologies, there’s one critical thing to keep in mind: When the contract for inclusion in the anthology was signed. If it was on or after 01 January 1978, there’s a high probability that the publisher’s claim for 100% because it was supposedly a “work made for hire” is incorrect, and instead the author(s) would share in whatever share belongs to the author(s) (see 2 above). This requires examination of the contract.
4. Also, if the contract assigns a different proportion of relevant rights to the author than “50% that includes e-books”, that may change things.
5. Perhaps most importantly: This is a proposed settlement. Those with somewhat-more-than-short-term memories may recall the Google Book Search attempted settlements… which first time was sent back for revisions, then rejected in its entirety. There is a new judge handling the Anthropic case since the end of December, and she has already signalled that There Will Be Issues to Resolve; indeed, she has specifically stated in an order that the hearing on 23 April that was supposed to consider whether the settlement can be approved will consider only attorney’s fee motions, and that the “fairness hearing” will be rescheduled for a later date. (Indeed, the motions and briefs to approve the settlement aren’t complete yet, and the fairness hearing cannot be held less than 60 days after that.)
Disclaimer: None of the above is legal advice. I have not appeared in any case cited above, nor do I forseeably intend to do so. Those works of mine that might otherwise fall inside this settlement class were properly classified as “work made for hire” (they’re all “educational”), so I have no direct conflict of interest regarding this settlement. I here express no opinion on the adequacy, propriety, or other merits of this settlement — only explain some of the “who’s entitled to what” aspects that are clear to me from the settlement itself (as an experienced class-action and copyright-matters litigator) that have escaped much public comment.
Even more important disclaimer: If you think you need legal advice on this, get it from an attorney experienced with either copyright matters or class-action settlements. Not from social media, any other part of the internet, or someone you run across randomly at a convention. Not from any “internet advocacy” person/organization. Not from “claims processing” firms, like (but not only) the one excoriated by the previous judge in this case (in November 2025). Not from “writers’ groups,” ranging from small internet-based interest groups up to and including big national organizations that claim to speak for “all authors.” Get your own legal advice for your own circumstances.
dwsmith
Thank you, C.E. and I keep harping on writers to know copyright. This kind of observations from someone like you who KNOWS copyright can do nothing but help writers understand how little they really know about how they make their money in publishing. So thank you!!!
Ashley Pollard
Oh wow! What a rip off by Anthropic. Damn them all to hell. Hope you get some restitution, or at least satisfaction in nailing them by adding your works to the claim.
dwsmith
What makes me shake my head are the writers who know they are using stolen material and still use AI in their writing and art. And then thinking they own it.
What is funny about Anthropic is that they stole stolen material. The others, like Chat GPT bought a copy and then thought they could use that copy for anything they wanted.
LM
I remember back when that lawsuit was made, the legal analysis iirc was they did not train their AI on it. Some enterprising employee intended to and the company shelved it.
The fact that they pirated it is why they owe everyone. They didn’t even have to use it.
Which is a big reason I don’t pirate things, no matter the justifications the average person makes for the practice.
Vincent Zandri
Thats a major chunk of change if they do pay Dean. Like you said, it’s worth the effort. I believe we’ll get something.
dwsmith
I believe we will to, when and how much is the question. (grin)
Jason Morgan
Stunning move by Anthopic. They calculated that stealing the books and paying the subsequent fine would be outweighed by the first-mover advantages of getting their tool to market fast and powerfully. They simply couldn’t wait for the copyright office if they wanted to become a major player in their field.
dwsmith
Nah, no thought of copyright or breaking the law and they honestly didn’t care if they had thought of it.
C.E. Petit
Dean, I have to disagree. It’s worse than that.
Many of these idjits believe — because this is the way they want it to be, and therefore it is — that information wants to be free, and everything is just information.
Don’t laugh too loud. This is precisely the attitude I was tangling with in Ellison v. AOL, and especially from those who thought Harlan (and I) wanted to destroy the still-developing Internet. And we didn’t even have to deal with social media yet, over a quarter of a century ago!
They’re in trouble with the first part of their aphorism. Not all “pure information” wants to be free (presuming it has volition in the first place); trade secrets, classified documents, the details of your financial accounts… the list goes on. Where they really shoot themselves in much more delicate places than “the foot,” though, is the second part. “Original expression” is not “just information.” That’s part of the difference between, say, Shakespeare’s Sonnet XIV (“Not from the stars…”) and the factual statement “Astronomy and astrology do not predict anything in human affairs, but some people believe they do and they’re useful literary devices.”
And that’s why these cretins didn’t care: They did not believe that there was anything that actually mattered that was not freely available to them. “The legal consequences” were… meaningless information. Any disparagement of this method of thinking by comparison to that of doomsday cults, “primitive” magical systems, and the moral methods of two-year-olds is purely intentional. And the techbros criticize us for being obsessed with “speculative fiction”…