Artificial Intelligence and copyright: a happy (or even possible) relationship?

This content has been archived. It may no longer be relevant

From The 1709 Blog:

[I]n the realm of IP one of the questions that have been asked with increasing frequency is whether and to what extent AI has the potential to replace humans, including in the creative fields.

As AI machines become increasingly autonomous, can they be regarded as ‘authors’ in a copyright sense and, if so, can the works they create be eligible for copyright protection? If the answer was again in the affirmative, who would own the copyright in such works?

. . . .

For instance, readers with an interest in music might have had the opportunity to listen to the recently released single Hello Shadow, which is the first song extracted from the the first multi-artist music album composed with AI.

This album was curated by Benoit Carré, head of SKYGGE, who collaborated with several musicians and performers, including – in the case of Hello Shadow – Stromae and Kiesza.

The SKYGGE project started as a research project (the Flow-Machines project, conducted at Sony Computer Science Laboratories and University Paris 6) in which scientists were looking for algorithms to capture and reproduce musical “style” [an example being Daddy’s cara song in the style of the Beatles]. However, the novelty and huge potential of the approach triggered the attention of musicians who joined the team.

It is clear that SKYGGE produces music thanks to AI, but there is a substantial human input. But as things have the potential to develop in the sense that AI will be able to create music entirely on its own, without any human input, will the resulting songs be protected by copyright?

. . . .

[A]t the international level there is no definition of who is to be regarded as an ‘author’ in a copyright sense. However, legal scholarship seems oriented in the sense of concluding that, from its text and historical context, under the Berne Convention only natural persons who created the work can be regarded as authors.

In any case, although generally speaking it seems possible “to agree that an author is a human being who exercises subjective judgment in composing the work and who controls its execution”, this does not mean that at the national level there are not situations in which also works created by non-human authors can qualify for protection, or courts have not addressed issues of non-human authorship.

. . . .

Harmonization of the standard of originality at the EU level has been limited. Only the Software Directive (Article 1(3)), the Database Directive (Article 3(1)) and the Term Directive (Article 6) provide that, respectively, for computer programs, databases and photographs copyright protection shall be only available if they are their “author’s own intellectual creation”.

. . . .

One may wonder how a non-human author can exercise such rights. The question becomes even more complex, if not impossible to solve, if one considers that the CJEU has clarified that the language of that directive imposes that authors are considered as the exclusive first owners of economic rights.

Link to the rest at The 1709 Blog

PG just checked the number of posts he has created for TPV.

The number is . . . . . 16,130.

PG stifled an OCD impulse to figure out a way to determine how many sentences and words are contained in those 16,130 posts.

PG doesn’t think MS Word is up to the task, but he’s never tried. (OCD stifling is going quite well at the moment.)

However, PG wondered if he could create a short program to rearrange segments of the 16,130 posts to create an almost infinite number of new posts and transform TPV into a perpetual blog.

(PG understands that perpetual motion devices are supposed to be impossible, but, since blog posts don’t contain motion, physics may not negate the concept of a perpetual blog. PG will stop trying to think about physics now.)

PG couldn’t stop wondering whether MS Word could be up to some sort of word-counting task for those 16,130 blog posts (OCD scores one point). Quite frankly, he had his doubts about MS Word.

So, he copied and pasted 20 copies of the ms from Mrs. PG’s forthcoming novel into a single MS Word file, expecting the word-counting function to fail.

PG apologizes to all the programmers in Redmond.

20 copies of Mrs. PG’s next book contain:

  • 2081 pages (MS Word manuscript pages, not pages in a printed book or ebook pages)
  • 43,500 paragraphs
  • 1,154,740 words
  • 6,295,780 characters (with spaces)

As PG looks at these numbers, he may need to retract his apologies.

Other than the page count, MS Word says that 1) the number of paragraphs, 2) the number of words and 3) the number of characters in 20 copies of Mrs. PG’s next book are each divisible by 10.

What are the chances of this happening? (PG is becoming tired of numbers, so he won’t go down that path).

Ever dubious, PG suspected a bit of fudging by the programmer responsible for the word count feature in MS Word. Was Microsoft just guessing?

He was about to make an accusation.

But first, he added a single character to this monstrous file.

(pregnant pause)

(no sexism intended)

And . . .

the characters count increased by 1. To a number not divisible by 10.

Then he entered a space and another single character.

And the word count increased by 1 while the character count increased by 2 (one space and one character).

Then he hit the Enter key and another single character.

And the paragraph count increased by 1 with character and word count each incrementing properly.

PG can confirm that the paragraph, word and character counting algorithms in his copy of MS Word appear to be accurate (No, he’s not going to personally count words. He’s an attorney, not an accountant.), at least for a document containing 20 copies of Mrs. PG’s next book.

He unconditionally withdraws any and all disparaging comments directed towards any and all nameless programmers in Redmond.

19 thoughts on “Artificial Intelligence and copyright: a happy (or even possible) relationship?”

  1. “the number of characters in 20 copies of Mrs. PG’s next book are each divisible by 10.”

    If you multiply the number of words/characters/paragraphs by 20, the divisible by 10 is pretty inevitable.

    • What can I say. I didn’t have time to explore all the permutations & combinations & implications & railstations, Gary.

  2. The problem with trying to teach an A.I. how to make something a human will like is that we ourselves still don’t understand humans enough to know what they’ll like.

    So I don’t expect to see any AI stories coming out (other than as jokes), but I can see an AI replacing people like Patterson(?) – framing a few ideas to base a story around. (I’ve seen another writer using random numbers to not use the same generic ‘body’ over and over. Does M/F matter to the story? If ‘no’ then flip a coin. Race? Roll a die; same for hair color/length/style/clothes.)

  3. I recently had reason to look up the maximum file size possible in Word 2016. Apparently it’s a wee bit more than 500 MB. That’s over half a gigabyte. I humbly suggest you might double Mrs PG’s book count and still have change. 🙂

  4. I haven’t tried to stress Word 2016 much, but I was involved in a group that brought 2010 to its knees. It was a technical standards working group. We were using a template with a helaciously complex style-sheet and Word 2010 started falling apart at around 200MB, as I remember.

    Luckily, one of the members of the working group was a Msft employee who was on the Word project team. He took our issue up with his colleagues, who suggested simplifying the template. The combination of a complex template combined with larger file sizes took us into lightly tested territory.

    Fortunately for fiction authors and most non-fiction authors, most of the templates they use have well less than a hundred styles and a simple structure. Non-fiction templates tend to be more complex, but I haven’t seen anything close to the monstrosity we were using.

    The template we were using to meet international standards formatting rules had many hundreds of styles and very complex rules for structuring with sub-sub-sub-… styles. The author(s) of the template had done some stuff that the Msft QA guys had not even dreamt about– I often wonder about the dreams of a software test engineer– and had baked some inconsistencies into the template. It was not too hard to straighten the mess out.

    The lesson is that if you think Word is choking on file size, examine the style sheet for your template before drawing your seppuku sword.

    Complete aside: I’ve noticed that complex templates slow Word performance. Another reason to favor simple templates.

    • I keep simplifying the templates/styles I use to format ebooks because of the limits of what a lot of ereaders can output to their screens.

      • Agree. I am pleased with the appearance of my CreateSpace paper book formatted with a simple Word template of my devising, but the Kindle version, meh. Good, but not distinctive. The fine points are all lost.

        But I still prefer reading on the Kindle.

        On the other hand, I find that a Word doc that looks good on Kindle, doesn’t meet CreateSpace requirements. In the final reckoning, formatting for CreateSpace first has been less total work for me.

  5. My recent experience suggests that almost all of the plain text editors on the market will choke on a manuscript of that length.

    My test corpus was Clarissa by Samuel Richardson, which is well over 900,000 words. It’s one of the longest books in the English language — the longest one I could find that had public domain text available online.

    TextMate, Sublime Text, Ace Editor, Microsoft’s Monaco, and the (unnamed) editor I was running comparisons against all worked well (no crashes or slowdowns). Most of the others… did not. Some simply went catatonic. Others worked, sort of, but with a several second delay for each keystroke. Find and replace was an “enter the command, then go to lunch” kind of thing.

    It’s good to know that Word can handle a task of that size. I didn’t test word processors, just text editors, but I would not be surprised if many word processing programs will also become unusably slow (or crash outright) when fed that much text.

    • I am curious– did you load Clarissa into Word? I started to try it myself just now, but gave up when I saw that I would have to merge 9 vols from Gutenberg. Easy, but not easy enough for idle curiosity. It still can’t be much more than 6 or 8 MB of text.

      I can’t think of a large docx that I’ve worked with that was not bulked up with graphics. I have a suspicion Msft has not put much new effort into text handling since XP days.

      • > I am curious– did you load Clarissa into Word?

        Nope. I was only testing plain text editors, not word processors.

        > It still can’t be much more than 6 or 8 MB of text.

        5.2 MB merged, not counting the Project Gutenberg boilerplate.

Comments are closed.