<
https://www.techdirt.com/2023/10/05/publishing-a-book-means-no-longer-having-control-over-how-others-feel-about-it-or-how-theyre-inspired-by-it-and-that-includes-ai/>
"There’s no way to write this article without some people yelling angrily at
me, so I’m just going to highlight that point up front: many, many people are
going to disagree with this article, and I’m going to get called all sorts of
names. I actually avoided commenting on this topic because I wasn’t sure it was
worth the hassle. But I do think it’s important to discuss and I’ve now had two
separate conversations with authors saying they agree with my stance on this,
but are afraid of saying so publicly.
I completely understand why some authors are extremely upset about finding out
that their works were used to train AI. It feels wrong. It feels exploitive. (I
do not understand their lawsuits, because I think they’re very much confused
about how copyright law works. )
But, to me, many of the complaints about this amount to a similar discussion to
ones we’ve had in the past, regarding concerns about if works were released
without copyright, what would happen if someone “bad” reused them. This sort of
thought experiment is silly, because once a work is released and enters the
messy real world, it’s entirely possible for things to happen that the original
creator disagrees with or hates. Someone can interpret the work in ridiculous
ways. Or it can inspire bad people to do bad things. Or any of a long list of
other possibilities.
The original author has the right to speak up about the bad things, or to
denounce the bad people, but the simple fact is that once you’ve released a
work into the world, the original author no longer has control over how that
work is used and interpreted by the world. Releasing a work into the world is
an act of losing control over that work and what others can do in response to
it. Or how or why others are inspired by it.
But, when it comes to the AI fights, many are insisting that they want to do
exactly that around AI, and much of this came to a head recently when
The
Atlantic released a tool that allowed anyone to search to see which authors
were included in the Books3 dataset (one of multiple collections of books that
have been used to train AI). This lead to a lot of people (both authors and
non-authors) screaming about the evils of AI, and about how wrong it was that
such books were included.
But, again, that’s the nature of releasing a work to the public. People read
it. Machines might also read it. And they might use what they learn in that
work to do something else. And you might like that and you might not, but it’s
not really your call.
That’s why I was happy to see Ian Bogost publish an article explaining why he’s
happy that his books were found in Books3, saying what those two other authors
I spoke to wouldn’t say publicly. Ian is getting screamed at all over social
media for this article, with most of it apparently based on the title and not
on the substance. But it’s worth reading."
Cheers,
*** Xanni ***
--
mailto:xanni@xanadu.net Andrew Pam
http://xanadu.com.au/ Chief Scientist, Xanadu
https://glasswings.com.au/ Partner, Glass Wings
https://sericyb.com.au/ Manager, Serious Cybernetics
