Unravelling the bolero

Maurice Ravel (1875-1937) died without descendants. He left eighty-five completed works, of which the one everybody knows is Boléro. “On his death the copyright passed to Ravel’s brother, who married his caretaker after a car accident. She then left the rights to her first husband, who married his manicurist. Ultimately the copyright wound up belonging to the manicurist’s daughter from a previous marriage, Evelyne Pen de Castel.” (The Economist, May 4th, 2024.) Allegedly Boléro is played by some orchestra somewhere in the world on average every fifteen minutes or so, so this inheritance was a pretty valuable gift. Royalties over the years are estimated to have topped $100 million.

However as with all good things, the clock eventually runs out. Under French copyright law Boléro entered the public domain in 2016.* But two years later here comes a law suit by the Benois family claiming that as set designer for the ballet in which the piece was first performed, Alexandre Benois, was a co-creator of the work and should thus be regarded as co-owner of the copyright. Because another alleged co-creator, Bronislava Nijinska, the choreographer, didn’t die until 1972, royalties they claim should continue to flow (to the co-authors’ descendants too) until 2053. Ms Pen de Castel has unsurprisingly joined this suit — a third of $100 million is better than zilch! A decision from the court in Nanterre is expected in June.

One might assume, along with SACEM’s† deputy chief executive, David El Sayegh, that the lawsuit is just a bit of nonsense. The notion of co-authors is surely a stretch. The genesis of the piece was a commission in 1922 from dancer Ida Rubinstein for an orchestration of six piano pieces by Issac Albéniz. When Ravel discovered that the music had already been orchestrated and was protected by copyright, he switched to composing his own work, based on the Spanish bolero dance form. The ballet piece premiered at the Paris Opera in November 1928. Boléro was published in 1929 by Durand publishers. As well as the orchestral version, arrangements were published for piano, both solo and duet, and later, in 1930, for two pianos.

Let us hope that the court agrees that the suit is nonsensical. If it were to succeed this might appear to condemn us to claims to royalties from the typesetter of a book, the printer, the binder, the truck driver, and more realistically I suppose, the copyeditor, the editor, and the publicity manager who persuaded that newspaper editor to review the book resulting in a review that sent hordes to their bookstores demanding copies. The reviewer would of course also have a good claim to have contributed to the success of the project.

________________

* In almost all countries Boléro has already moved into the public domain, but in USA it remains under copyright until 1 January 2025.

†  Société des auteurs, compositeurs et éditeurs de musique

Abridgment?

At The Bookshelf: Editions and Subtractions on the website Public Discourse, Matthew Frank considers whether abridgment strikes at the integrity of a book. (Link via Today in Books.) You can perfectly legitimately summarize a book’s plot/content without quoting (much of) it, but an abridgment can’t be published without the permission of the copyright holder. What would make an author want to do an abridgment of their book? Or allow someone else to do it? Money, I suppose; or more politely, the possibility of reaching another group of potential readers. After all if you thought your novel would really be better if it was shorter, wouldn’t you have written it that way to begin with?

To my mind the classic instance of abridgment is the Readers Digest Condensed Books series (since 1997 known as Reader’s Digest Select Editions).

Photo: Terriloui. Wikimedia Commons.

I was somewhat overwhelmed to visit almost fifty years ago one of R. R. Donnelley’s plants in Chicago which was devoted entirely to the manufacturing of this series — there must have been millions of books lying around in process through the large plant: all ready to feed a voracious appetite for reading the latest book without actually reading all of the latest book. The series started in 1950 when four volumes a year (it’s now six) were sent to subscribers. Each volume contains abridgments of four (occasionally more) books of popular fiction and non-fiction.

The most recent volume listed in the Wikipedia article, is Number 397, issued in January 2024. It contains

• The Lie Maker by Linwood Barclay
• An Evil Heart – Linda Castillo
• Same Time Next Summer – Annabel Monaghan
• Flop Dead Gorgeous – David Rosenfelt

Each volume costs $24.99, plus $6.99 shipping, but for your $32 you are getting four books. In 1987 the New York Times estimated that annual sales were ten million copies, though I dare say it’s now a bit less. Still profitable enough to continue though.

Ten million readers getting twenty-four books a year: not too shabby! Does it really matter that the books were abridged? You can appreciate them, enjoy them, and get more than enough out of the abridgments to discuss the books at cocktail parties — and reading condensed books is certainly better than reading no books.

Many a foreign book has been introduced to a local audience in abbreviated form. Is it better to have a hint of Hafiz or to remain in ignorance?

Alexander Bubb has an interesting piece at the OUP blog about this question, focussed mainly on the Rubaiyat of Omar Khayyam. We see, almost as a dilettante display of “culture”, Edward Fitzgerald passing the time translating these ruba’i (quatrains) from Persian while sauntering around the Sussex fields near his home. I confess to being inclined to the absolutist party in this question. I rather suspect that Louisa Stuart Costello’s 1845 anthology of lyric poets, The Rose Garden of Persia, muddied the waters rather than did service to international understanding. Still, just writing that sentence pulls me up short — that can’t really be, can it? It’s almost the same élitist impulse (the same one that makes me put an acute accent on that word!) as wanting to read a poet in his/her Complete Works of . . . , rather than in the Selected Poems of X, or in say Palgrave’s Golden Treasury. Isn’t it just snobbery? Obviously to some extent the poet intended that you should read this poem after reading that one, or they’d not be in that sequence, but reading this poem and never reading that one has to be better than never reading either. Many an early translator seems to have regarded their role as an editor as well as translator, ignoring passages they didn’t like (or maybe had difficulty translating?). A full draught of vintage may be better than a few sips, but a taste has to be better than no wine at all.

‘Scholarly Press Backlist Revival’

Orphan books are copyright books whose author cannot be traced, and so are largely unusable for extensive quotation, adaptation and other protected uses since there’s nobody to ask for permission. A few years ago one bright idea for how to spring the orphans from their captivity was described in Orphanage. This appears not to have worked: no doubt it got tangled up in the Internet Archive lawsuit.

Lost along with the orphans are lots of old academic volumes with known authors whose books ceased years ago to command sales sufficient to justify their being kept in print. Via Kathy Sandler’s Technology · Innovation · Publishing comes news that New York Public Library believes it has come up with a new method to save all those not-quite-orphaned forgotten books (at least those from scholarly presses) from the oblivion the law has inadvertently condemned them to. H-Net carries the story.

“’Scholarly Press Backlist Revival’, our project provides an alternative to tracking down individual contracts. The key — what we believe will allow us to scale this effort but remain well within the bounds of what we believe is fair and based on collaboration and consent — is asking presses to grant NYPL permission to make their books digitally available to the world not on a title-by-title basis, but on entire out-of-print backlist basis. For those works publishers aren’t sure they still hold the rights to, we will also go to authors to seek their permission.”

Previous attempts by NYPL to activate this inaccessible category of books have foundered on the inability of authors to be sure whether they actually have the right to give permission for NYL to allow their customers access to out-of-print books, allied to the lack of financial benefit to publishers who’d have to incur the costs of checking old contracts. NYPL plans to walk past this impasse by getting permission from both the publishers, on a group basis, i.e. permission covering all their out of print books, as well as from the authors.

A little niggle at the back of my mind carries the label “print-on-demand”. “Out-of-print backlist” sounds like something publishers would have no problem giving up on a list-wide basis, but — is there really anything called “out-of-print backlist” any longer? Any book can be set up for POD — as long as the publisher still has the rights. Of course the books NYPL is talking about are by definition OP books, but just being OP doesn’t mean always that rights have been reverted, especially in the academic world where the publisher often acquires copyright ownership. So if the thought enters the mind of the publisher faced with the NYPL deal that maybe they could sell the book as a POD book . . .

If such a thought arises, let’s hope that the next step in the thought process is taken. Let’s do both: make sure the NYPL lease specifies digital rights only. We all know, I think, that having a book available electronically at libraries is liable to increase demand not decrease it. Not by much I dare say, but we are dealing with small numbers here — which may be what stymies us once again. Maybe it’s just too expensive to have to iron out any kind of complexity.

Copyright law changes . . . please!

Oh God, what a mess!

At The New Yorker Louis Menand has a long review of Who owns this sentence? by David Bellos and Alexandre Montague.* Copyright’s completely out of control. It’s being pulled in so many directions that it’s easy to forget that it was first established “to promote the Progress of Science and useful Arts, by securing for limited Times to Authors and Inventors the exclusive Right to their respective Writings and Discoveries.” The shenanigans Mr Menand describes around pop music and graphic art just don’t seem to be in the same universe as the idea of protecting The Uptake and Storage of Noradrenaline, and encouraging researchers to keep on with the good work. Actually of course, even in the narrow restricted world of academic research, the “promoting of science and the useful arts” bit is itself subject to a slight ambiguity — after all these guys are paid a salary to teach and do research, so isn’t that enough to keep them at it? And to make their work product a work made for hire for that matter? (Let’s leave that can of worms firmly closed.)

Mr Menand tells us “Virtually every song that Bruce Springsteen has ever written is now owned by Sony, which is reported to have paid five hundred and fifty million dollars for the catalogue. Because the copyright clock does not start ticking until the demise of the creator, Sony could own those rights until past the end of the century. The longer the Boss lives, the richer Sony gets.” Oh Citizens United — did we really have to accept that a corporation was a person?

Now, it goes without saying that in 1790 the technological options for someone wishing to promote science and the useful arts were a little narrower than they are today. I think this matters. Having an out-of-date law of copyright in an age of lightning-fast technological change is surely a huge problem. It ends up being, similarly to a common law regime, a matter of interpretation by the courts.  For almost anyone, taking a crack at someone else’s copyright is worthwhile — after all you might win. Just because this court ruled one way doesn’t have to mean that court won’t be persuaded to go in the opposite direction. Confusion results, and you can feel anything’s worth a try. A few years back Rod Stewart was sued for copyright infringement because he used a photo of the back of his head on a poster advertising a Las Vegas show. A photographer who had once photographed the back of Stewart’s head, and been lucky enough to have it used on a Stewart album cover, sued for copyright infringement because the idea was too close to their copyrighted image (not the one used of course). Louis Menand discussed this example of crazy law in a New Yorker piece entitled Copywrong in 2014.

We last redid copyright law in 1976, with another partial rejig in 1998, and the longer we leave discussion of revisions the harder the task becomes as the world keeps on changing faster and faster. Now we’ve got ChatGPT et al to deal with too. The good news (I hope it ends up being good news) is that the Copyright Office at the Library of Congress is actively thinking about the issue. The New York Times has an article about how the Copyright Office is studying what the implications for copyright are of all the new technologies which have arrived since the last review of the law, including but not restricted to Artificial Intelligence. (Link via LitHub.)

Seven years ago I was calling for us to split copyright into three different strands. It now seems that we may need more. Copyright is a nice simple idea — encouraging the search for knowledge by ensuring that its creators are rewarded, but now copyright has exploded far beyond that idea. Clearly all the things covered today by copyright do need to be protected, but they should probably be protected by different laws, leaving boring old copyright to boring old books. I could even see an argument for a separate type of protection for trade books and academic work.

But it’s not only about protecting: people need to have access too though, or the search for knowledge is just hampered, not facilitated. The Texas Law Review has a review of fair use law for a provocative proposal that “fair learning” should become the concept enabling machine learning to take place free of copyright restrictions. As the authors, Mark Lesley and Bryan Casey point out “It’s shockingly easy to create a copyrighted work. Copyrights cover a broad swath of creations, from the written word to art of all types to software, dance, and even architecture. . . . The standard for establishing copyright protection is low — you need only have an ‘original work of authorship’ and record it in some more-than-transitory form. . . . And both of those requirements are so trivial as to be almost meaningless.” As I write this, it becomes copyright as I save it. If I draw a doodle on a paper napkin, that’s copyright. Nobody cares; nobody wants it; and it’s valueless since I’m not Picasso — but for a reading/learning robot such a thing is valuable. How is a self-respecting machine to distinguish between my doodle and a David Hockney drawing?

It is so “shockingly easy” that in a response to the scale of the “problem” of AI-generated books Amazon has been forced to limit “authors” to publishing only three books a day! (Story via Kathy Sandler’s Technology · Innovation · Publishing.) Amazon/Kindle are pretty aggressive in their definition of the AI problem: “We define AI-generated content as text, images, or translations created by an AI-based tool. If you used an AI-based tool to create the actual content (whether text, images, or translations), it is considered ‘AI-generated,’ even if you applied substantial edits afterwards.” This seems overkill to me: if “substantial edits” equals a complete rewrite, surely that makes it your work.

We need to be getting our minds around the issues that call out for copyright review.† The Copyright Alliance has a webpage about AI and Copyright, which they will be continuously updating. (Link via Technology · Innovation · Publishing.) But it is important not to forget the other items which were causing concern before we were carpet bombed by AI.

______________________________

* This book, Who Owns This Sentence? A History of Copyrights and Wrongs, published by W. W. Norton at $28.99 does seem to demand attention. (Bookshop.com will give you a couple of bucks off.)

† But we need to remember that copyright law doesn’t affect just our jurisdiction; it’s an international system, and it behooves us to look at transnational implications of any law revision. Michael Healy of the Copyright Clearance Center, pointed out in 2016 that we should be careful what we wish for: not all changes are changes for the better.

AI authorship

Not sure who The New Publishing Standard has been listening to, but I’ve not been conscious that “authors, publishers and other creatives have been saying . . . that AI will never be able to replicate human creativity and quality.” I dare say, just as conservative anti-maskers would assert that Covid was fake news, that there’s been the odd blast of conversational bravado on the topic, but by and large I think the reaction has been that AI presents a huge and uncertain potential. The divergence, as far as I can see, is between those who think it is potentially a great thing, and those who see it as a potential Armageddon. But doesn’t everyone think it’s either a huge danger or a great potential benefit — or maybe both — not that’s it’s going to fizzle out into irrelevance and just go away?

The occasion for Mark Williams’ post is the announcement by Rie Kudan, winner of the biannual Akutagawa Prize, that her novel, Tokyo-to Dojo-to, had been written with the help of an AI Chatbot, and that indeed about 5% of it was direct quotation from the AI output. Mr Williams notes that there is as yet no information on whether this revelation will affect the Japanese copyright of the book, and he goes on to wonder whether this will get in the way of translations. But surely if a translator who’s a human, translates text that’s AI generated, we end up in the target language with words that were written by a human being, so such copyright worries are surely misplaced. Indeed it occurs to me that, if governments do dig their heels in and refuse copyright to anything with AI content, as seems to be their current stance, this presents the royal road to dodging this very copyright hang up. Get the Chatbot to write your novel in French, then translate it into English, and there you go. There’s no copyright in ideas, only in the expression of ideas, and here all the expression would be human-generated in English. This is precisely analogous with a writer editing material generated by AI in their native language, or even editing/changing stuff read elsewhere. (Of course all bets are off if the temptation to use Google Translate is acceded to.)

I have occasionally been taken to task for my apparent permissiveness in the face of Chatbots — but it’s not that I don’t think there are bad things that bad actors can do. Only yesterday we were hearing about a round of deepfake robot phone calls from “virtual Joe Biden”, attempting to depress turnout in that day’s New Hampshire Primary, — though do note the fake was promptly identified — and also about faked porno videos of schoolgirls in New Jersey. Of course any technology is capable of bad things — even pencils have occasionally generated pornography, but we are I think all aware of the need for some sort of regulation of AI. However I just believe there are lots of good things that will come from this technology, and that the good will outweigh the bad. Of course vigilance is the price of liberty, and we will have to work out methods of vigilance — with which I dare say AI will help us. When I’m reading a book why should I worry that this fictional insight into the human condition, or that advice suggestion may have been generated by a person or a machine. If the story’s good, and the advice works, isn’t that all that matters? God knows there are lots of human beings who can be relied upon to give you lousy advice, so why fear ChatGPT or its more tightly controlled successors?

Open AI, creators of ChatGPT, are, unsurprisingly, making an effort to stave off election damage resulting from their product. The Verge (link via Kathy Sandler’s Technology · Innovation · Publishing) tells the story. The most significant safeguard is the introduction of a certification system — though of course in order not to be fooled you’ll need to have paid attention to that. I suspect there’ll still be plenty of work for “virtual Joe”. (In a slightly different sense of “virtual”, “virtual Joe” did pretty well in NH yesterday, winning a sort of virtual primary as a write-in candidate!)

As far as copyright goes I do not think any book generated wholly by AI should be given copyright protection — the law was made to protect the earning power of people, not machines. However I think it’s about as sensible to deny copyright protection to a human author because the book was written with the help of AI, as it is to do so because they made use of a Thesaurus. If you think copyright protection should be denied to anyone using machine intelligence, why don’t you also think that anyone who writes with the assistance of a word-processing computer should be denied copyright protection? — After all earlier authors had to write things out by hand, so obviously that’s what the original copyright law was intended to protect!

See also AI & ©, and AI and copyright.

False © claims

Apparently copyright takeover is becoming a bit of a problem in the music industry. The New York Times tells us about Bad Dog, a Washington, DC folk duo who wanted to compile a give-away CD of their songs for a party. They’d been sharing their songs on-line via SoundCloud for ages without any desire to commercialize them, but when they came to create a CD they discovered a problem. By chance the producer of their album posted one of their songs, “Preston”, to his studio’s Instagram account, and was puzzled as to why Instagram labelled the song as “Drunk the Wine” by Vinay Jonge. Further investigation of Bad Dog’s songs revealed that “Pop song” had become “With me tonight” by Kyro Schellen, “The Misfit” had become “Outlier” by Arend Grootveld, “Verona” had become “I told you” by Ferdinand Eising and so on. None of these supposed artists had any other songs credited to them, nor any identity on the web. The company about to press Bad Dog’s CD ran a copyright check and found that Bad Dog didn’t even own copyright in these songs any longer — songs which they themselves had written over the previous years. This had been taken by whoever uploaded the songs.

One of the duo’s members is a practicing lawyer involved in intellectual property rights, while the other is a retired law professor who specialized in internet copyright, but even they found sorting the problem out almost impossible. In the end they sort of succeeded because they kept sending out “take down” notices. Apple Music responded to one with a form letter, and although that letter did little else at least by good fortune it included the name of the company which had been uploading Bad Dog’s music — it was Warner Music! It turns out that Warner owns a subsidiary, Level, a music distributor catering to independent artists. For a $20 annual subscription you too can be a musician — all you have to do is click a box agreeing to Level’s terms of service (which include a promise not to upload any audio owned by anyone else), and then fame awaits. Obviously nobody checks up on this ownership claim. Apparently Bad Dog’s songs were played about 60,000 times on Spotify, and this would have yielded income of about $250 at Spotify’s rates. Not a fortune, but “This is a scalable scam, said Mr Batey of Beatdapp [a company involved in tracking this sort of fraud]. SoundCloud boasts more than 320 million songs, many of them the work of weekend noodlers. These people may never realize that their work has been grabbed and renamed and is syphoning money from the royalty pool.” As one of the members of Bad Dog says “I couldn’t get a deal with Warner to save my life. But they made money from my music, and that money was from straight-up infringement”.

Moving to the book business, Plagiarism Today brings us an article entitled 5 Warning Signs a Copyright Notice May be False. I assumed that when they say copyright notice they are referring to the copyright notice printed in the book — but they are not. They are talking about fake take down notices! The article shows however that this sort of false-ownership copyright fraud is not restricted to music — books are getting to play in this muddy puddle too. If you just write without any thought of money this may not be such a problem — after all if your material turns out to be available under a different nom de plume, so what? It’s almost a compliment. If however there’s the slightest chance that at some remote date in the future you might wish to publish this material, or even just be recognized for having created it, (maybe even quote from it!) it might behoove you to exercise some vigilance. Just what that would mean is unfortunately not altogether obvious. Clearly if you are self-publishing you are a bit further down this lane of risk than someone who’s just writing on-line. You’ve already demonstrated that your writing is worth something.

I guess a musician might be advised to set up a Level account and get their music out there under their own name. A writer, though? Presumably you’ve not become a self-publisher because you didn’t want the hassle. Is hassle eventually necessary?

Salinger and ©

Olivia Rutigliano asks at LitHub “Did J.D. Salinger Wield Copyright as Self-Protection?”

The answer is complicated but seems ultimately to be a kind of “Yes”. It may appear almost perverse to publish a book which is so wildly successful that you can afford to buy a hideaway where you aim to protect yourself from the attentions of the public, so that they can never ask you about the book. The Catcher in the Rye (1951) is one of those bestseller-of-all-time candidates. Salinger’s “Hapworth 16, 1924”, which was published in The New Yorker in 1965 was his last public writing. He just wanted to be left alone, and came to regard publication as a sort of invasion of privacy: “Suppose you had a coat you liked,” he told the Times in 1974, “and somebody went into your closet and stole it. That’s how I feel.”

The struggle to be left alone provoked the outside world into efforts to thwart him. “Already at loggerheads with publishers and agents, he later added reviewers, academics, journalists and biographers to his extensive list of enemies, taking on the outside world most dramatically in 1986 when he sought to block Ian Hamilton’s biography, In Search of JD Salinger, in the courts.” (— The Guardian Eighty Years of Solitude.) The fact that Holden Caulfield seems to have inspired a couple of assassins might be thought to complicate matters. “Copyright protections can stop a work from being copied, pirated, poached. They can’t stop it from being misunderstood.” For a man who just wanted to be left alone Salinger was rather litigious, suing to stop Ian Hamilton’s biography as well as two sequels to The Catcher in the Rye. He won the Hamilton biography lawsuit on the basis that unpublished letters could not be the basis for a claim of fair use — they argued that the copyright laws required that the material “fairly” used had to have already been published. One consequence of this restriction on fair use copying was that the University of Maryland felt unable to copy Katherine Anne Porter’s unpublished personal papers before the paper they were written on decayed away and much was lost to posterity. (The law has subsequently been adjusted in this regard.)

J. D. Salinger, at the age of 69, trying to punch the camera out of a photographer’s hands in 1988. Photo: New York Post, Paul Adao

Fame/celebrity is an obvious double-edged sword — so many people buy your book; so you can afford to hide yourself away in an effort to avoid them. But once you’ve published the bestseller, the damn thing’s passed beyond your control. Basking in the glory of dumb-assed celebrity, or hiding yourself away behind high walls, are surely not the only alternatives. Refusing all interviews just increases the desire to interview you. Getting testy about the invasions of your privacy brings about doesn’t help: people don’t react well to a grumpy old man trying to smash their cameras. Pynchon decamped to Mexico. Harper Lee seems to have managed better. Shakespeare should have told us about a celebrity who hates celebrity — or is that Lear? Probably the odd public appearance would have yielded more privacy than a total embargo on interviews. Salinger’s literary heirs — the estate is currently run by his widow, Colleen, and his son, Matt — are no doubt as copyright-protective as was the author, though they did not contest a biography by Kenneth Slawenski. Salinger’s books should surely be in The Library of America series. Is the fact that he’s not there due to the estate? No idea. The Catcher in the Rye is protected by copyright until 2046*. No doubt lots of royalties by then — though being published also in a handsome hardback collection might be unlikely to impact paperback sales.

_______________

* The mathematically/legally curious may be asking why that’s not 2080. After all Salinger died in 2010 and the current rule of “life + 70 years” should apply, shouldn’t it? However all of Salinger’s works were published before the Copyright Act of 1976 introduced this provision. In 1951, the expectation was 28 years plus one renewal of a further 28 years. The 1976 Act granted copyrights already in force on 1 January 1978 (when the law came into effect) a total term of 95 years from the date of publication (or registration in the case of unpublished works). 1951 + 95 = 2046.

AI and law

At last a lawyerly look at the most recent panicky question in our book publishing lives: what about the copyright status of AI, both coming (training) and going (outputs)? Publishers Weekly brings us a sober analysis of what’s at stake. it’s author, Lloyd J. Jassin is a publishing attorney, former publishing executive, and coauthor of The Copyright Permission & Libel Handbook, published by John Wiley in 2010.

Mr Jassin downplays the dangers. “Copyright law” he reminds us “strikes a balance between those who create content and the public’s interest in having wide access to that content. It does this via granting authors a limited monopoly over the dissemination of original works by giving them the exclusive right to reproduce, distribute, and create derivative works based on copyrighted material.” Common sense kind of makes you reflect that this cannot really apply to computers and their output.

Reinforcing my belief that this new technology will enhance the role of the gatekeeper in our industry, Mr Jassin concludes “In an age of disinformation, an author’s brand, a publisher’s imprint, and the goodwill associated with them are valuable assets. I believe the industry is less vulnerable than many think. But, to quote Nick Lowe, ‘Where it’s goin’ no one knows’.”

Here’s a piece from the ABA Journal by Erik Sherman balancing the issues. (Link via Kathy Sandler’s Technology · Innovation · Publishing.) It’s true the article is subtitled “What a Mess”. They see the issue as having three parts: input, algorithms, and output. Input is perhaps not an infringement: after all “copying into memory is not copyright infringement.” Nevertheless, The New York Times is off to law on this. Copying to create an algorithm can be viewed as fair use, with the algorithm as a derivative work — which just introduces a different argument. Output can be problematic because in theory a chatbot could reproduce entire passages of the original training work, but this is difficult to cause, so difficult to prove. Identifying liability may also be a problem: for example OpenAI apparently requires the protection of full indemnification as part of their “terms of use”.  

But copyright seems almost to be the least of our problems. The facility with which “bad actors” can generate plausible AI-assisted nonsense is terrifying. Deepfakes are now absurdly easy to create: here’s Ethan Mollick demonstrating the point — apart from the very beginning of this video everything in it is fake, made up by AI software. He doesn’t speak Hindi. (From One Useful Thing.)

If you don’t see a video here, please click on the title of this post in order to view it in your browser.

During this election, if you hear Joe Biden saying something which sounds scandalous, do not believe it unless you see him saying it yourself. Deepfakes will be flying back and forth. Clearly it won’t just be politics — a convincing scientific paper may be simply concocted, with lots of authoritative-looking back up. Mistrust becomes a virtue. This sort of skepticism will ultimately have to be our defense against such stuff — unless it comes from a trusted gatekeeper anything controversial you see online should probably be treated as a lie. Too bad: gone is the dream of a free internet bringing all humankind together for good. I wonder if we can try again? Or do we just have to retreat into the analog world?

See also AI & ©.

Legal deposit video

If you don’t see a video here, please click on the title of this post in order to view it in your browser.

I wrote about legal deposit in October, just a few days before David Crotty at The Scholarly Kitchen delivered this video showing the system’s consequences. Featured is the British Library’s immense remote storage facility in Boston Spa. Rather amazing.

It’s not just books that they keep. They store all sorts of ephemera, even little handbills warning about the onset of Covid. “This is the raw text of history as it happened, and someone has to keep it preserved for the future.” Not unfortunately if Valancourt Books have their way.

I did do a post about the British Library a few years ago — but the video about Boston Spa linked to there has been taken down by the BBC.

Public Domain Day, 2024

Duke University School of Law has a post covering this day in detail. “On January 1, 2024, thousands of copyrighted works from 1928 will enter the US public domain, along with sound recordings from 1923.” Among the notable books now in the public domain are the following:

(Funny how Lady C covers tend to feature a lady, usually clothed it’s true, but never a semi-clad gamekeeper wearing a deer-stalker with shotgun at the ready. It’s he who’s mentioned in the title after all.)

Also bouncing into public domain status today is Mickey Mouse, but only in his first manifestation. Steamboat Willie from 1928 is now available — all the films showing the character we’d recognize are still in copyright. Mickey himself is protected by a trademark defense. Disney’s tactics appear to have changed — they are now more exercised by online infringement, and seem to have called off the lobby dogs on copyright “reform”.

But why is it that books published in 1928 are the ones now crossing the public domain barrier? Isn’t the term of copyright life + 70 years? The answer lies in the Mickey Mouse Copyright Act of 1998. Duke explains “The 1998 Copyright Term Extension Act gave works published from 1923 through 1977 a 95-year term, expiring on January 1 after the conclusion of the 95th year. Doing the math, works from 1928 were copyrighted for 95 years—through 2023—and are in the public domain January 1 2024.” So for the rest of (most of) our lives get ready for this transition to make no sense at all.

But look out for the copyright timing in translations. Im Westen Nights Neues may be in the public domain today, but All Quiet on the Western Front isn’t. The first English translation by A. W. Wheen came out in 1929, so hold off for twelve months unless your adaptation/parody is in German. I see that the cover I chose to show is the cover of a graphic novel adaptation — no, no, don’t copy that version: it’s still copyright. Another oddity is Peter Pan which today enters the public domain in USA, but remains in perpetual copyright in Britain. Indeed nothing said here has any relevance in Britain, Europe and other copyright domains with their own terms of protection.

The Duke Law School post concludes with an impassioned discussion of the madness of our constantly extending terms of copyright. My wish list includes Copyright three ways.