Not everyone needs to have an opinion on AI
National Novel Writing Month is an American initiative that has become a worldwide pastime, in which participants attempt to write a 50,000-word manuscript in the month of November. Some of these first drafts eventually become novels — the initial version of what became Erin Morgenstern’s The Night Circus started life as a NaNoWriMo effort — but most don’t. And many participants cheerfully admit they are writing for the pleasure of creation rather than out of any expectation that they will gain either money or prestige from the activity.
In recent years, NaNoWriMo has been plagued by controversies. This year, the organisation has been hit by an entirely self-made argument, after declaring that while it does not have an explicit position on the use of generative artificial intelligence in writing, it believes that to “categorically condemn the use of AI writing tools” is both “ableist and classist”. (The implication that working-class people and people with disabilities can only write fiction with the help of generative AI, however, is apparently A-OK.)
The resulting blowback saw one of its board, the writer Daniel José Older, resign in disgust. (NaNoWriMo has since apologised for, and retracted, its initial statement.)
There is very little at stake when you participate in NaNoWriMo, other, perhaps, than the goodwill of the friends and relations you might ask to read your work afterwards. Sign-ups on the website can talk to other participants on their discussion forums and are rewarded for hitting certain milestones with little graphics marking their achievement. If you want to write an experimental novel called A Mid-Career Academic’s Reflections Upon His Divorce that is simply the same four-letter expletive repeated over and over again, nothing is stopping you from doing so. If you want to type the words “write the first 50,000 words of a coming-of-age novel in the style of Paul Beatty” into ChatGPT and submit the rest, you can do so. In both cases, it is your own time you are wasting.
The whole argument is exceptionally silly but does hold two useful lessons.
One is that organisations and companies should have fewer opinions. Quite why NaNoWriMo needs to have an opinion about the use of generative AI is beyond me. Organisations should have a social conscience, but that should be limited to things they actually directly control. They should care about fairness when hiring, about the effects that their supply chains have on the world, just as NaNoWriMo should care about whether its discussion forums are well moderated (the subject of another previous controversy). But they should have little or no interest in issues that they have no meaningful way to stop or prevent, like what participants do with AI.
A good rule of thumb for an organisation considering whether to make a statement about a topic is to ask itself what material changes within its control it proposes to make as a result of doing so — and why. Those changes might range from donating money to hiring. For example, the cosmetics retailer Lush has given large amounts of money to police reform charities, while Julian Richer, the founder of Richer Sounds, home entertainment chain, went so far as to turn his business into an employee-owned trust in 2019.
But if an organisation is either unwilling or incapable of making real changes to how it operates or spends money, then nine times out of ten that is an indication that it will gain very little and add very little from speaking out.
The second lesson concerns how organisations should respond to the widespread use and adoption of generative AI. Just as NaNoWriMo can’t stop me asking Google Gemini to write a roman-à-clef about a dashingly handsome columnist who solves crimes, employers can’t reliably stop someone from writing their cover letter by the same method. That doesn’t mean they should necessarily embrace it, but it does mean that some forms of assessment have, inevitably, become a test of your ability to work well with generative AI as much as your ability to write or to research independently. Hiring, already one of the most difficult things any organisation does, is already becoming more difficult, and probation periods will become more important as a result.
Both lessons have something in common: they are a reminder that organisations shouldn’t sweat the stuff outside of their control. Part of writing a good novel is choosing the right words in the right places at the right time. So too is knowing when it is time for an organisation to speak — and when it should stay silent.
Posting != Endorsing the writer’s views.
The reasonable in-between is despising without presently fearing.
GenAI is a plagiarism engine. That’s really not something that can be defended. But as a means of automating away the jobs of writers it has proven itself to be so deeply deficient that there’s very little to fear at this time.
The arrival of these tools has, however, served as a wake up call to groups like the screenwriters guild, and I’m very glad that they’re getting proper rules in place now before these tools become “good enough” to start producing the kind of low grade verbal slurry that Hollywood will happily accept.
Human artists / writers take influence from others as well. Nobody is creating art in a vacuum and I don’t see generative AI much different from the way humans operate. I’d argue it’s virtually impossible to write a sentence that has not been written before and every new human-created art piece probably has a really close equivalance that has already been done before.
These things sound analogous if you know very little about how both generative AI and the human creative process actually work.
I mean this sincerely: why bother getting excited about anything, then?
A new Marvel movie, a new game, a new book, a new song. If none of them are unique in any way, what is the point of it all? Why have generative AI go through this song and dance? Why have people do it? Why waste everyone’s time?
If the plagiarism engine is acceptable because it’s not possible to be unique anyway… I just, I don’t know how you go on living. It all sounds so unbelievably boring.
Just because you’re using standard materials, it doesn’t mean you can’t combine them in a unique ways and even if every possible sentence has been said before, that doesn’t mean everyone has heard it before too.
The point is that not being allowed to pull from existing content is an impossible standard. Nobody is being as original as they may think they are and even when you truly come up with an idea independently it’s highly likely that you’re not the first to think of that.
Why do you find it such a depressing idea? I face this attitude often when discussing free will as well, which I genuinely don’t believe in either but that has zero effect on my ability to get excited or motivated about things.
Okay, so you don’t believe new things can’t be unique. You just think that plagiarism is when one person uses the word ‘the’ and then a second person uses the word ‘the’.
That art is dead? Through sheer saturation alone, no one has anything left to say? That watching the new Cinderella is line-by-line the same as watching the old Cinderella, and the money machine keeps this corpse moving along only because people are too stupid to realize they’re being sold books from a library? I really don’t know how you couldn’t.
This is like asking me why a polluted lake is sad.
I’ll ignore the first part as it doesn’t represent my view.
I don’t think art is dead and I disagree with the implication that AI simply hands you a copy of something somebody else did before. That’s not how generative AI works. There would be nothing generative about that. Instead it studies the prior work of humans, finds patterns and combines these in unique and novel ways. If I ask for a hybrid of Harry Potter and Lord of the rings then obviously its using existing building blocks but the outcome is still something that has not been written before.
I’m an artist myself. I take photographs. I’m under no illusion that all my photos are completely unique, they’re not. I’m well aware that if I had a database of every single picture ever taken, then there would hundreds if not even thousands of photos that are near identical to the ones I’ve taken and are so proud of. That takes zero joy out of my creative process and from the enjoyment other people find in my work. Nobody has seen every photo in the world. My art is not meaningless just because someone did it before me.
I was equivocating singular words and entire sentences on purpose.
If you can recombine sentences in interesting ways, into paragraphs that are your own ideas, that isn’t plagiarism. Why would “people can’t construct unique sentences either” be a rebuttal if that’s not what plagiarsm is?
You’re anthropomorphising.
LLMs are little clink-clink machines that produce the most typical output. That’s how they’re trained. Ten thousand inputs say this image is of a streetlight? That’s how it knows.
The fact an LLM knows what a Lord of Rings is at all means that Tolkien’s words, the images, the sounds, are all encoded in its weights somewhere. You can’t see them, it’s a black box, but they live there.
Could you say the same of the human brain? Sure. I know what a neuron is.
But, LLMs are not people.
All of that is besides the point, though. I was just floored by how cynical you could be about your own supposed craft.
A photograph of, say, a pretty flower is fantastic. As an enjoyer of art myself, I love it when people communicate things. People can share in the beauty that you saw. They can talk about it. Talk about how the colors and the framing make them feel. But if you’re view is that you’re not actually adding anything, you’re just doing more of what already exists, I really don’t know why you bother.
Okay, assume someone has. Is your art meaningless, then? All of photography is just spectacle, and all the spectacles have been seen?
That’s not my view.
No. That’s my point. Art isn’t dead.
The plagiarism engine effect is exactly what you need for a good programming tool. Most problems you’re ever going to encounter are solved and GenAI becomes a very complex code autocomplete.
An LLM constructed only out of open source data could do an excellent job as a tool in this capacity. No theft required.
For writing prose it’s absolutely trash, and everyone using it for that purpose should feel ashamed.
In theory, yes, although having actually tried using genAI as a programming tool, thy actual results are deeply lacklustre at best. It sort of works, under the right circumstances, but only if you already know enough to confidently do the job yourself, at which point the value in having an AI do it for you, and then having to check the AI’s work for any of a million possible fuck ups, seems limited at best.
Yeah my usage of it is similarly limited. But the plagiarism engine is more useful than it is annoying in my experience. Especially in writing kdoc or unit test variations. Write one, write the name of the next, have autocomplete fill it out with the expected conditional variation
I feel like you could probably create a helper script to do that more reliably.
Ish.
You’ll have assertions that are entirely new or different, other pieces of setup or teardown. It really is one of the best use cases for GH’s Copilot that I’ve run across.
In my day to day the intellij autocomplete is what I prefer.
Noted. I’ll have to play around with that sometime.
Despite my obvious stance as an AI skeptic, I have no problem with putting it to use in places where it can be used effectively (and ethically). I’ve just found that in practice, those uses are varnishingly few. I’m not on some noble quest to rid the world of computers, I just don’t like being sold overhyped crap.
I’m also hesitant to try to rebuild any part of my workflow around the current generation of these tools, when they obviously aren’t going to exist in a few years, or will exist but at an exorbitant price. The cost to run genAI is far, far higher than any entity (even Microsoft) has any willingness to sustain long term. We’re in the “give it away or make it super cheap to get everyone bought in” phase right now, but the enshittification will come hard and fast on this one, much sooner than anyone thinks. OpenAI are literally burning billions just in compute right now. It’s unsustainable. Short of some kind of magical innovation that brings those compute costs down a hundred or thousand fold, this isn’t going to stick around.
Oh I’m 1000% in agreement with you. I think Copilot for programming is more expensive than it’s worth right now, both for my employer and for Microsoft.
OpenAI et al have done nothing to address the fundamental issue of hallucinations. In code hallucinations are pretty quickly evident: your IDE immediately throws up error highlights whenever the code complete fucks up.
The latest open AI model is to chain together a computational centipede to try and create reasoning structures out of stochastic processes. It takes longer and still doesn’t fix the issues. In their own demo video there are clear bugs with the “code” their 4o model writes.